Communicating Results to Your Staff

10/12/22, 1:51 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken=… 1/45

Part IV Best Practices in Healthcare Analytics Across the Ecosystem

10/12/22, 1:51 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken=… 2/45

17 Overview of Healthcare Analytics Best Practices Across the Ecosystem

Dwight McNeill

Analytics in healthcare is old and new. Science has been a strong underpinning of healthcare in the research devoted to the discovery of causes and treatments of disease. However, delivering this knowledge from the bench to the bedside to optimize the care of every patient has been an ongoing challenge. Although treating sickness, that is, the interaction between a patient and her caregivers, is the raison d’être of healthcare, the industry is more complex than that. The American way of healthcare requires large doses of payment, finance, regulation, research and development, and administrative and business supports. Healthcare is a huge part of the U.S. economy, accounting for 18% of GDP at a spending rate of $8,500 for every

man, woman, and child.1 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch17#ch17end01) And it is big business. Annual hospital expenses are approaching $1 trillion, and physician services are more than $0.5 trillion. Both of these categories of providers amount to more than 50% of spending. The next highest spending area is for prescription drugs, which amounts to 10% of spending.

Healthcare is both an informational and a personal business. It is personal because it deals with people, and communications and relationship skills are fundamental to making change happen. It is informational in that it is about discovery, measurement, improvement, and running a business.

Analytics is the high octane fuel to feed the thirsty information engines. It holds the promise to improve people’s lives, increase revenues and reduce costs, and to change the very nature of what healthcare is and what it can be.

Part IV (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/part04#part04) is on best practices and includes eight case studies of leading organizations in healthcare analytics. These are bellwether organizations and represent the best of the art and science of analytics as of 2012. The case studies are inclusive of the settings where analytics is practiced including providers, payers, and a life sciences company. It includes both the public and private sectors.

The chapters in part IV (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/part04#part04) address the “whats” and “hows” of analytics to support organizational strategies and goals. The whats include the domains of the content of analytics, including clinical, business, and marketing purposes. The hows include the functions of analytics, including how it is organized, how it adds value, and its technical challenges.

Providers Providers are the boots on the ground in healthcare: the doctors, nurses, and a myriad of other care professionals who interact with patients and their families to treat them. Their efforts are important to people’s well-being, and very often spell the difference between life and death. This is where the rubber meets the road, where the clinical knowledge created by research, the pills and devices developed by life science companies, the financial coverage provided by insurance companies, the healthcare benefits provided by employers, and the support by all the business functions of hospitals and other healthcare settings all come into play. As such it is a vitally important fulcrum for analytics to support clinicians with information, knowledge, and the tools to improve practice.

The providers included in the case studies are very large, ranging in revenues from just under $1 billion to over $150 billion. They are integrated delivery systems that provide a continuum of care from hospital to outpatient care in an organized and coordinated way and are accountable for the populations they serve both clinically and fiscally. Because they are accountable, they are more incentivized to use analytics and make continual improvements.

The Whats

These five best practice providers in Part IV (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/part04#part04) , including Partners Healthcare System, Catholic Health Initiatives, Veterans Health Administration, Air Force Medical Service, and HealthEast Care Sytstem, have been at the vanguard of health analytics partly because of four common areas of content and focus:

• They were early adopters of electronic health records (EHRs). EHRs support their patient care strategies, such as care coordination, disease management, and use of care protocols, by increasing the availability of individual patient and population data and by improving communication among providers. Partners Healthcare, the VA, and the Air Force had EHRs in place systemwide by the early 1990s. Note that

only 35% of U.S. hospitals had adopted EHRs by 2011.2 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch17#ch17end02) Early wins from EHRs included Computerized Provider Order Entry (CPOE) systems, which improved the accuracy of physicians. medication orders, and also measuring adherence to medication guidelines. For example, adverse drug events were cut in half after the introduction of these systems at a Partners hospital.

• The leadership for clinical analytics is clear at these organizations. They want to achieve clearly articulated institutional goals such as reducing medical errors, achieving uniformly high clinical quality, improving chronic disease management, and using clinical resources efficiently. One of the keys to the transformation of the VA was a performance measurement system that was used to hold senior managers accountable for improvement in performance measures. The analytics undergirding the accountability system include tracking metrics and reporting on them through dashboards. Similarly, HealthEast set out on a “quality journey” to become the benchmark for quality in the Twin Cities area and deployed analytics for measurement and improvement strategies.

• They use a clinical data warehouse for research purposes. For example, Partners uses it for postmarket surveillance to detect problems with drugs and medical devices after they are released to the market. The VA detected an outbreak of a rare form of pneumonia and was able to determine that a certain nasal spray was the cause. The warehouse also provides the data foundation for supporting many forms of research, which can garner big revenues for these institutions.

• Finally, these institutions use analytics for business and finance functions including optimizing revenue, understanding employer attrition, claims adjudication, and reimbursing physicians based on performance metrics such as cost-effective use of imaging services.

10/12/22, 1:51 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken=… 3/45

The Hows

Much of the work in analytics in healthcare today is building capacity specifically related to connecting the data “pipes” and integrating the data including various forms of clinical, operational, and financial data. The Partners case study demonstrates the issues involved in deciding what goes into an enterprise level analytics design versus a hospital-specific design. Similarly, the overarching question of Chapter 19 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch19#ch19) , “Catholic Health Initiatives,” is how does a healthcare organization translate data into actionable information for every stakeholder across the enterprise? The need for an efficient and scalable data warehouse is discussed in Chapter 21 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch21#ch21) , “The Health Service Data Warehouse Project at the Air Force Medical Service (AFMS) (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch21#ch21) .” It addresses the challenges of finding, acquiring, improving, and integrating data and reducing long lead times and frustration on the part of users. Chapter 22 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch22#ch22) , “Developing Enterprise Analytics at HealthEast Care System (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch22#ch22) ,” focuses on how to organize analytic teams at different levels to accomplish different purposes.

Payers Payers, including a multitude of commercial health insurers, employers, and governments, provide the financing for the high cost of health services in the United States. Payers face epic challenges, including the advent of health information exchanges, health insurance exchanges, new Affordable Care Act (ACA) regulations on coverage, premium reviews, profit margins and mandates, new provider models such as Accountable Care Organizations (ACOs), and a huge new pool of customers who were previously uninsured. Payers also face the dual demands to 1) change their business model of providing wholesale insurance to employers to providing retail health and insurance services to individuals while also 2) focusing on the health and management of populations.

Payers had gotten into a routine of managing the economics of benefits and coverage, premium pricing, and various insurance products, but were not as actively engaged in managing health and medical care of members/employees as they were for the brief but noteworthy managed care era of the 1990s. Now, the tenor has changed and the pendulum has swung back and beyond such that insurers are changing the very nature of their business by blurring the lines between payers and providers to ensure better business results and also by changing their mission to become a health company and/or to become an information company where insurance is just one product line for these organizations.

The analytics challenges and opportunities are daunting. Payers have relied on claims data as their intelligence source to understand their business but will need to rely on diverse data to address the above challenges. This diverse data coupled with need to comply with unrelenting regulations will necessitate the review of legacy systems, the capacity of existing data warehouses, and a heightened need to integrate, process faster, discover insights, and contribute regularly to the bottom line.

There are two chapters in Part IV (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/part04#part04) on payers, including a health insurer—Chapter 23 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch23#ch23) , “Aetna (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch23#ch23) ,” by Kyle Cheek—and an employer—Chapter 24 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch24#ch24) , “Employee Health and Benefits Management at EMC: An Information Driven Model for Engaged and Accountable Care (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch24#ch24) ,” by David Dimond and Robert Morison.

Cheek notes that Aetna’s analytic maturity is high (4+) on the Davenport-Harris Analytics Maturity Model relative to an industry average of stage 2. He describes the five primary services for its internal and external constituencies including 1) provider analytics to identify opportunities for outcome and cost improvements among physicians and hospitals, 2) plan-sponsor reporting for employers, 3) program evaluation of the ongoing effectiveness of care management programs, 4) custom informatics for “special projects,” and 5) data warehousing.

In terms of the factors of analytics success, Cheek says that the most important are identifying the strategic drivers that offer the most demonstrable value from analytical enhancement, lodging the data warehouse with the informatics organization, and developing an internal analytics competency.

Diamond and Morison describe a different analytics focus, on the employee, and how the company promotes health for its workforce. The EMC vision is for employers to engage patients and providers, enable health awareness and literacy, influence health and lifestyle behaviors, and drive adoption of patient-centric technologies. The analytics to support the employee focus include an employee health portal, a personal health record, health risk assessments and incentives to be healthy, and the availability of related health management programs.

Life Science Companies As Handelsman stated in Chapter 4 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch04#ch04) , “Surveying the Analytical Landscape in Life Sciences Organizations (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch04#ch04) ,” the life sciences industries are engaged in discovering, developing, and commercializing new therapies. The industry challenges are no less daunting than the rest of the health ecosystem. These include blockbuster drugs going off patent, health plan pressures to lower costs and demanding justification for the value of drugs, pricing pressures from generic drugs, long lasting investor caution following the Great Recession, the high cost and failure rate of clinical trials, and cost cutting that has cut into research innovation.

Analytics have been a core skill in the research and development discovery process and in determining value through comparative effectiveness studies. But, as in other aspects of healthcare, putting the research knowledge to use in improving clinical and business outcomes has lagged. There is great promise with the analytics of personalized medicine and the use of genomic to fill in gaps in products. And there is a renewed focus on customers that goes beyond direct to consumer advertising that can build loyalty and foster brand support.

In Chapter 25 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch25#ch25) , “Commercial Analytics Relationships and Culture at Merck (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch25#ch25) ,” Davenport reports on one specific life science industry analytical

10/12/22, 1:51 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken=… 4/45

function, commercial analytics that focus on promotion and sales, at a major pharmaceutical firm, Merck. He concentrates on the “how” of making analytics work well for the business. It’s all about decision support for the sales business. According to the business, the analytics function has been successful because the group members are “thought partners”: they start with a full understanding of the business question and then marshal data to answer the questions, they are “field friendly” in translating findings into solutions, and they embed analytical results into software tools. Key questions about the future role of analytics are how to expand beyond the U.S. market and provide global support and how to create more collaboration among other analytics groups at Merck.

Notes 1 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch17#ch17end01a) . Micah Hartman et al., National Health Spending in 2011:

Overall Growth Remains Low, but Some Payers and Services Show Signs of Acceleration, Health Affairs 32 (2013): 87-99.

2 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch17#ch17end02a) . U.S. Department of Health & Human Services, “HHS Secretary Kathleen Sebelius Announces Major Progress in Doctors, Hospital Use of Health Information Technology,” February 12, 2012, www.hhs.gov/news/press/2012pres/02/20120217a.html (http://www.hhs.gov/news/press/2012pres/02/20120217a.html) .

10/12/22, 1:51 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken=… 5/45

18 Partners HealthCare System

Thomas H. Davenport

Partners HealthCare System (Partners) is the single largest provider of healthcare in the Boston area. It consists of 12 hospitals, with more than 7,000 affiliated physicians. It has 4 million outpatient visits and 160,000 inpatient admissions a year. Partners is a nonprofit organization with almost $8 billion in revenues, and it spends more than $1 billion per year on biomedical research. It is a major teaching affiliate of Harvard Medical School.

Partners is known as a “system,” but it maintains substantial autonomy at each of its member hospitals. While some information systems (the electronic medical record, for example) are standardized across Partners, other systems and data, such as patient scheduling, are specific to particular hospitals. Analytical activities also take place both at the centralized Partners level and at individual hospitals such as Massachusetts General Hospital (MGH) and Brigham and Women’s Hospital (usually described as “the Brigham”). In this chapter, both centralized and hospital-specific analytical resources are described. The focus for hospital-specific analytics is the two major teaching hospitals of Partners—MGH and the Brigham—although other Partners hospitals also have their own analytical capabilities and systems.

Centralized Data and Systems at Partners The basis of any hospital’s clinical information systems is the clinical data repository, which contains information on all patients, their conditions, and the treatments they have received. The inpatient clinical data repository for Partners was initially implemented at the Brigham during the 1980s. Richard Nesson, the Brigham and Women’s CEO, and John Glaser, the hospital’s chief information officer, initiated an outpatient electronic medical

record (EMR) at the Brigham in 1989.1 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch18#ch18end01) This EMR contributed outpatient data to the clinical data repository. The hospital was one of the first to embark on an EMR, though MGH had begun to develop one of the first full-function EMRs as early as 1976.

A clinical data repository provides the basic data about patients. Glaser and Nesson came to agree that in addition to a repository and an outpatient EMR, the Brigham—and Partners after 1994, when Glaser became its first CIO—needed facilities for doctors to input online orders for drugs, tests, and other treatments. Online ordering (called CPOE, or Computerized Provider Order Entry) would not only solve the time-honored problem of interpreting poor physician handwriting, but could also, if endowed with a bit of intelligence, check whether a particular order made sense or not for a particular patient. Did a prescribed drug comply with best-known medical practice, and did the patient have any adverse reactions in the past to it? Had the same test been prescribed six times before with no apparent benefit? Was the specialist to whom a patient was being referred covered by his or her health plan? With this type of medical and administrative knowledge built into the system, dangerous and time-consuming errors could be prevented. The Brigham embarked on its CPOE system in 1989.

Nesson and Glaser knew that there were other approaches to reducing medical error than CPOE. Some provider institutions, such as Intermountain Healthcare in Utah, were focused on close adherence by physicians to well-established medical protocols. Others, like Kaiser Permanente in California and the Cleveland Clinic, combined insurance and medical practices in ways that incented all providers to work jointly on behalf of patients. Nesson and Glaser admired those approaches, but felt that their impact would be less in an academic medical center such as Partners, where physicians were somewhat autonomous, and individual departments prided themselves on their separate reputations for research and practice innovations. Common, intelligent systems seemed like the best way to improve patient care at Partners.

In 1994, when the Brigham and Mass General combined as Partners HealthCare System, there was still considerable autonomy for individual hospitals in the combined organization. However, from the onset of the merger, the two hospitals agreed to use a common outpatient EMR called the longitudinal medical record (LMR) and a CPOE system, both of which were developed at the Brigham. This was powerful testimony in favor of the LMR and CPOE systems, since there was considerable rivalry between the two hospitals, and Mass General had its own EMR.

Perhaps the greatest challenge was in getting the extended network of Partners-affiliated physicians up on the LMR and CPOE. The physician network of more than 6,000 practicing generalist and specialist physician groups was scattered around the Boston metropolitan area, and often operated out of their own private offices. Many lacked the IT or telecom infrastructures to implement the systems on their own, and implementation of an outpatient EMR cost about $25,000 per physician. Yet full use of the system across Partners-affiliated providers was critical to a seamless patient experience across the organization.

Glaser and the Partners information systems (IS) organization worked diligently to spread the LMR and CPOE to the growing number of Partners hospitals and to Partners-affiliated physicians and medical practices. To assist in bringing physicians outside the hospitals on board, Partners negotiated payment schedules with insurance companies that rewarded physicians for supplying the kind of information available from the LMR and CPOE. By 2007, 90% of Partners-affiliated physicians were using the systems, and by 2009, 100% were. By 2009, more than 1,000 orders per hour were being entered through the CPOE system across Partners.

The combination of the LMR and the CPOE proved to be a powerful one in helping to avoid medical error. Adverse drug events, or the use of the wrong drug for the condition or one that caused an allergic reaction in the patient, typically were encountered by about 14 of every 1,000 inpatients. At the Brigham before LMR and CPOE, the number was about 11. After the widespread implementation of these systems at Brigham and Women’s, there were just above five adverse drug events per 1,000 inpatients—a 55% reduction.

Managing Clinical Informatics and Knowledge at Partners The Clinical Informatics Research & Development (CIRD) group, headed by Blackford Middleton, is one of the key centralized resources for healthcare analytics at Partners. Many of CIRD’s staff, like Middleton, have multiple advanced degrees; Middleton has an MD, a Master of Public Health degree, and a Master of Science in Health Services Research.

The mission of CIRD is

10/12/22, 1:51 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken=… 6/45

to improve the quality and efficiency of care for patients at Partners HealthCare System by assuring that the most advanced current

knowledge about medical informatics (clinical computing) is incorporated into clinical information systems at Partners HealthCare.2

(http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch18#ch18end02)

CIRD is part of the Partners IS organization. It was CIRD’s role to help create the strategy for how Partners used information systems in patient care, and to develop both production systems capabilities and pilot projects that employ informatics and analytics. CIRD’s work had played a substantial role in making Partners a worldwide leader in the use of data, analysis, and computerized knowledge to improve patient care. CIRD also has had several projects funded by U.S. government health agencies to adapt some of the same tools and approaches it developed for Partners to the broader healthcare system.

One key function of CIRD was to manage clinical knowledge, and translate healthcare research findings into daily medical practice at Partners. In addition to facilitating adoption of the LMR and CPOE, Partners faced a major challenge in getting control of the clinical knowledge that was made available to care providers through these and other systems. The “intelligent CPOE” strategy demanded that knowledge be online, accessible, and easily updated so that it could be referenced by and presented to care providers in real-time interactions with patients. There were, of course, a variety of other online knowledge tools, such as medical literature searching, available to Partners personnel; in total they were referred to as the “Partners Handbook.” At one point after use of the CPOE had become widespread at Brigham and Women’s, a comparison was made between online usage of the Handbook and usage of the knowledge base from order entry. There were more than 13,000 daily accesses through the CPOE system at the Brigham alone, and only 3,000 daily accesses of the Handbook by all Partners personnel at all hospitals. Therefore, there was an ongoing effort to ensure that as much high-quality knowledge as possible made it into the CPOE.

The problem with knowledge at Partners was not that there wasn’t enough of it; indeed, the various hospitals, labs, departments, and individuals were overflowing with knowledge. The problem was how to manage it. At one point, Tonya Hongsermeier, a physician with an MBA degree who was charged with managing knowledge at Partners, counted the number of places around Partners where there was some form of rule-based knowledge about clinical practice that was not centrally managed. She found about 23,000 of them. The knowledge was contained in a variety of formats: paper documents, computer “screen shots,” process flow diagrams, references, and data or reports on clinical outcomes—all in a variety of locations, and only rarely shared.

Hongsermeier set out to create a “knowledge engineering and management” factory that would capture the knowledge at Partners, put it in a common format and central repository, and make it available for CPOE and other online systems. This required not only a new computer system for holding the thousands of rules that constituted the knowledge, but an extensive human system for gathering, certifying, and maintaining the knowledge. It consisted of the following roles and organizations:

• A set of committees of senior physicians who oversaw clinical practice in various areas, such as the Partners Drug Therapy Committee, which reviewed and sanctioned the knowledge as correct or best known practice

• A group of subject matter experts who, using online collaboration systems, debated and refined knowledge such as the best drug for treating high cholesterol under various conditions, or the best treatment protocol for diabetes patients

• A cadre of “knowledge editors” who took the approved knowledge from these groups and put it into a rule-based form that would be accepted by the online knowledge repository

High Performance Medicine at Partners Glaser and Partners IS had always had the support of senior Partners executives, but for the most part their involvement in the activities designed to build Partners’ informatics and analytics capabilities was limited to some of the hospitals and those physician practices that wanted to be on the leading edge. Then Jim Mongan moved from being president of MGH (a role he had occupied since 1996, shortly after the creation of Partners) to being CEO of Partners overall in January 2003. Not since Dick Nesson had Glaser had such a strong partner in the executive suite.

Mongan had come to appreciate the value of the LMR and CPOE, and other clinical systems, while he headed Mass General. But when he came into the Partners CEO role, with responsibility over a variety of diverse and autonomous institutions, he began to view it differently. Mongan said:

So when I was preparing to make the move to Partners, I began to think about what makes a health system. One of the keys that would unite us was the electronic record. I saw it as the connective tissue, the thing we had in common, that could help us get a handle on utilization, quality, and other issues.

Together Mongan and Glaser agreed that while Partners already had strong clinical systems and knowledge management compared to other institutions, a number of weaknesses still needed to be addressed (most importantly that the systems were not universally used across Partners care settings), and steps needed to be taken to get to the next level of capability. Working with other clinical leaders at Partners, they began to flesh out the vision for what came to be known as the High Performance Medicine (HPM) initiative, which took place between 2003 and 2009.

Glaser commented on the process the team followed to specify the details of the HPM initiative:

Shortly after he took the reins at Partners, however, Jim had a clear idea on where he wanted this to go. To help refine that vision, several of us went on a road trip, to learn from other highly integrated health systems such as Kaiser, Intermountain Healthcare, and the Veterans Administration about ways we might bring the components of our system closer together.

Mongan concluded:

We also were working with a core team of 15-20 clinical leaders and eventually came up with a list of seven or eight initiatives, which then needed to be prioritized. We did a “Survivor”-style voting process, to determine which initiatives to “kick off the island.” That narrowed down the list to five Signature Initiatives.

The five initiatives consisted of the following specific programs, each of which was addressed by its own team:

10/12/22, 1:51 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken=… 7/45

• Creating an IT infrastructure—Much of the initial work of this program had already been done; it consisted of the LMR and the CPOE, which was extended to the other hospitals and physician practices in the Partners network and maintained. This project also addressed patient data quality reporting, further enhancement of knowledge management processes, and a patient data portal to give patients access to their own health information.

• Enhancing patient safety—The team addressing patient safety issues focused on four specific projects: 1) providing decision support about what medications to administer in several key areas, including renal and geriatric dosing; 2) communicating “clinically significant test results,” particularly to physicians after their patients have left the hospital; 3) ensuring effective flow of information during patient care transitions and handoffs in hospitals and after discharge; 4) providing better decision support, patient education, and best practices and metrics for anticoagulation management.

• Uniform high quality—This team addressed quality improvement in the specific domains of hospital-based cardiac care, pneumonia, diabetes care, and smoking cessation; it employed both registries and decision support tools to do so.

• Chronic disease management—The team addressing disease management focused on prevention of hospital admission by identifying Partners patients who were at highest risk for hospitalization, and then developed health coaching programs to address patients with high levels of need, for example, heart failure patients; the team also pulled together a new database of information about patient wishes about end- of-life decisions.

• Clinical resource management—At Jim Mongan’s suggestion, this team focused on how to lower the usage of high-cost drugs and high-cost imaging services; it employed both “low-tech” methods (e.g., chart reviews) and “high-tech” approaches (e.g., a data warehouse making transparent physicians’ imaging behaviors relative to peers) to begin to make use of scarce resources more efficiently.

Overall, Partners spent about $100 million on HPM and related clinical systems initiatives, most of which were ultimately paid for by the Partners hospitals and physician practices that used them. To track progress, a Partners-wide report, called the HPM Close, was developed that shows current and trend performance on the achievement of quality, efficiency, and structural goals. The report was published quarterly to ensure timely feedback for measuring performance and supporting accountability across Partners.

New Analytical Challenges for Partners Partners had made substantial progress on many of the basic approaches to clinical analytics, but there were many other areas at the intersection of health and analytics that it could still address. One was the area of personalized genetic medicine—the idea that patients would someday receive specific therapies based on their genomic, proteomic, and metabolic information. Partners had created the i2b2 (Informatics for Integrating Biology and the Bedside), a National Center for Biomedical Computing that was funded by the National Institutes of Health. John Glaser was co-director of i2b2 and developed the IT infrastructure for the Partners Center for Personalized Genetic Medicine. One of the many issues these efforts addressed in personalized genetic medicine was how relevant genetic information would be included in the LMR.

Partners was also attempting to use clinical information for postmarket surveillance—the identification of problems with drugs and medical devices in patients after they have been released to the market. Some Partners researchers had identified dangerous side effects from certain drugs through analysis of LMR data. Specifically, research scientist John Brownstein’s analyses suggested that the level of patients with heart attack admissions to Mass General and the Brigham had increased 18% beginning in 2001 and returned to its baseline level in 2004, which coincided with the timeframe for the beginning and end of Vioxx prescriptions. Thus far the identification of problems had taken place only after researchers from other institutions had identified them, but Partners executives believed it had the ability to identify them at an earlier stage. The institution was collaborating with the Food and Drug Administration and the Department of Defense to accelerate the surveillance process. John Glaser noted:

I don’t know that we’ll get as much specificity as might be needed to really challenge whether a drug ought to be in a market, but I also think it’s fairly clear that you can be much faster and involve much fewer funds, frankly, to do what we would call the “canary in the mine”

approach.3 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch18#ch18end03)

Partners was also focused on the use of communications technologies to improve patient care. Its Center for Connected Health, headed by Dr. Joe Kvedar, developed one of the first physician-to-physician online consultation services in an academic medical setting. The Center was also exploring combinations of remote monitoring technologies, sensors (for example, pill boxes that know whether today’s dosage has been taken), and online communications and intelligence to improve patient adherence to medication regimes, engagement in personal health, and clinical outcomes.

In the clinical knowledge management area, Partners had done an impressive job of organizing and maintaining the many rules and knowledge bases that informed its “intelligent” CPOE system. However, it was apparent to Glaser, Blackford Middleton, and Tonya Hongsermeier—and her successor as head of knowledge management, Roberto Rocha—that it made little sense for each medical institution to develop its own knowledge base. Therefore, Partners was actively engaged in helping other institutions with the management of clinical knowledge. Middleton (the principal investigator), Hongsermeier, Rocha, and at least 13 other Partners employees were involved in a major Clinical Decision Support Consortium project funded by the U.S. Agency for Healthcare Research and Quality. The consortium involved a variety of other research institutions and healthcare companies, and was primarily focused on finding ways to make clinical knowledge widely available to healthcare providers through EMR and CPOE systems furnished by leading vendors.

Despite all these advances, not all Partners executives and physicians had fully bought into the vision of using smart information systems to improve patient care. Some found, for example, the LMR and CPOE to be invasive in the relationship of doctor and patient. A senior cardiologist at Brigham and Women’s, for example, argued in an interview [with the author] that:

I have a problem with the algorithmic approach to medicine. People end up making rote decisions that don’t fit the patient, and it can also be medically quite wasteful. I don’t have any choice here if I want to write prescriptions—virtually all of them are done online. But I must say that I am getting alert fatigue. Every time I write a prescription for nitroglycerine, I am given an alert that asks me to ensure that my patient isn’t on Viagra. Don’t you think I know that at this point? As for online treatment guidelines, I believe in them up to a point. But once something is in computerized guidelines it’s sacrosanct, whether or not the data are legitimate. Recommendations should be given with notification of how certain we are about them.... Maybe these things are more useful to some doctors than others. If you’re in a subspecialty

10/12/22, 1:51 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken=… 8/45

like cardiology you know it very well. But if you are an internist, you may have shallow knowledge, because you have to cover a wide variety of medical issues.

Many of the people involved in developing computer systems for patient care at Partners regarded these as valid concerns. “Alert fatigue,” for example, had been recognized as a problem within Blackford Middleton’s group for several years. They had tried to eliminate the more obvious alerts, and to make changes in the system to allow physicians to modify the types of alerts they received. There was a difficult line to draw, however, between saving physician attention and saving lives.

Centralized Business Analytics at Partners While much of the centralized analytical activity at Partners has been on the clinical side, the organization is also making progress on business analytics. The primary focus of these efforts is on financial reporting and analysis.

For several years, for example, Partners has employed an external “software as a service” tool to provide reporting on the organization’s revenue cycle. It has also developed several customized analytics applications in the areas of cash management, underpayments, bad debt reserves, and charge capture. These activities primarily took place in the Partners Revenue Finance function.

The Partners Information Systems organization is also increasing its focus on administrative and financial analytics. It is putting in place Compass, a common billing and administrative system, at all Partners hospitals. At the same time, Partners has created a set of standard processes for collecting, defining, and modifying financial and administrative data. Further, as one article put it:

At Partners, John Stone, corporate director for financial and administrative systems, is developing a corporate center of business analytics and business intelligence. Some 12 to 14 financial executives will oversee the center, define Partners’ strategy for data management, and determine data-related budget priorities. “Our analysts spend the majority of their time gathering, cleaning, and scrubbing administrative data and less time providing value-added analytics and insight into what the data is saying,” says Stone. “We want to flip that equation so our

analysts are spending more time producing a story that goes along with the data.”4

(http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch18#ch18end04)

Hospital-Specific Analytical Activities—Massachusetts General Hospital MGH, because it was a highly research-driven institution, had long focused primarily on clinical research and the resulting clinical informatics and analytics. In addition to the LMR and CPOE systems used by Partners overall, MGH researchers and staff have developed a number of IT tools to analyze and search clinical data, one of which was a tool that searched across multiple enterprise clinical systems, including the LMR.

While historically, the research, clinical, information systems, and the analytically focused business arms of MGH tended to operate in stove pipes, the challenges of an evolving healthcare landscape have forced a change in that paradigm. For instance, a strong current focus within MGH is on how to achieve federal “meaningful use” reimbursement for the organization’s expenditures on EMR. Because achieving meaningful use objectives is predicated on a high level of coordination among information systems, the physicians, and business intelligence, people like David Y. Ting, the associate medical director for Information Systems for MGH and Massachusetts General Physicians Organization, and Chris Hutchins, the director of Finance Systems and deputy CIO, are beginning to collaborate extensively.

The HITECH/ARRA criteria for Stage 1 EMR meaningful use prescribe 25 specific objectives to incentivize providers to adopt and use electronic health

records.5 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch18#ch18end05)

To raise the level of EMR use by all its providers, as well as to provide resources for the work needed to achieve that level, MGH has arrived at a novel funds distribution model. They determined that the physicians organization will reserve a portion of the pool of $44,000 per physician toward IT and analytics infrastructure, then distribute the remaining incentive payment across all providers, proportional to the amount of data a particular physician is charged with entering. An internal quality incentive program would serve as the distribution mechanism. So, for example, if you recorded demographics, vital signs, and smoking status for the requisite number of patients, you would receive 30% of the per-physician payment from the pool. If you fulfilled all ten quality measures, you would receive 100% of the payment from the pool. This encourages all physicians to contribute to the meaningful use program, but it also means that no physicians will receive the full amount of $44,000. The incentive from the federal government is up to $44,000 for each eligible provider who fulfills the meaningful use criteria. MGH has examined the objectives and broken them down into ten major pieces of patient data that physicians need to record in the EMR. However, many are not relevant for all of its physicians. For example, a primary care physician would logically enter such data as demographics, vital signs, and smoking status, but these would be less relevant for certain specialists to enter.

Clearly, such a complex quality incentive model requires an unprecedented level of analytics. Currently, Ting, Hutchins, and others at MGH are working to map the myriad clinical and finance data sources that are scattered among individual departments, exist at a hospital site level, or exist at the Partners enterprise level. Simultaneously, they must negotiate data governance agreements even among other Partners entities, to ensure that the requisite data feeds from sources within Partners and pertaining to MGH, but stored outside MGH’s physical data warehouses, are available for MGH analytics purposes.

MGH has some experience with reimbursement metrics based on physician behaviors, having used them in Partners Community HealthCare, Inc. (PCHI), its physician network in the Boston area. Physician incentives have been provided through PCHI on the basis of admission rates, cost-effective use of pharmacy and imaging services, and screening for particular diseases and conditions, such as diabetes. This was also the mechanism used to encourage the adoption of the LMR and CPOE systems by physicians. But MGH, like other providers, struggles with developing clear and transparent metrics across the institution that can help to drive awareness and new behaviors. If MGH could create broadly accessible metrics on individual physicians’ frequency of prescribing generic drugs, for example, it would undoubtedly drive MGH’s competitive physicians to excel in the rankings.

10/12/22, 1:51 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken=… 9/45

On the business side, MGH is trying to develop a broad set of capabilities in business intelligence and analytics. A Business Intelligence/Performance Management group has recently been created under the direction of Chris Hutchins, deputy CIO and director of finance systems for the Mass General Physicians Organization (MGPO). The group is generating reports on such financial and administrative topics as

• Billing efficiency, claims adjudication, rejection rates, and times to resolve billing accounts, both at MGH overall and across practices

• Improving patient access, average wait times to see a physician, and cancellation and no show rates

• Employer attrition as an MGH customer

MGH is also working with CMS on the Physician Quality Reporting Initiative. To combine all these measures in a meaningful fashion, MGPO is also

working on a balanced scorecard.6 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch18#ch18end06)

While the current analytical activity is largely around reporting, Hutchins plans to develop more capabilities around alerts, exception reporting, and predictive models. The MGH Physicians Organization is implementing capabilities for statistical and predictive analytics that would be applied to several topics. For example, one key area in which better prediction would be useful involves patient volume. They are also pursuing more general models that would predict shifts in business over time. At the moment, however, Hutchins feels that the scorecard is still early in its development and current efforts are focused on identifying leading indicators.

Hospital-Specific Analytical Activities—Brigham and Women’s Hospital Like MGH, the Brigham’s analytical activities in the past have been largely focused on clinical research. Today it is also addressing much of the same business, operational, and meaningful use issues that MGH is. Many of the analytical activities at the Brigham are pursued by the Center for Clinical Excellence (CCE), which was founded by Dr. Michael Gustafson in 2001. The center has five functionally interrelated sections, including

• Quality programs

• Patient safety

• Performance improvement

• Decision support systems (including all internal and external data management and reporting activities)

• Analysis and planning (which oversees business plan development, ROI assessments for major investments, cost benchmarking, asset utilization reporting, and support for strategic planning)

The CCE has close working relationships with the Brigham’s CFO and finance organizations, the Brigham’s information systems organization, the Partners Business Development and Planning function, and other centers and medical departments at the Brigham.

One major difference between the Brigham and MGH (and most other hospitals, for that matter) is that the Brigham established a balanced scorecard beginning in 2000. It was based on a well-established cultural orientation to operational and quality metrics throughout the hospital. Richard Nesson, the Brigham CEO who had partnered with CIO John Glaser to introduce the LMR and CPOE systems, was also a strong advocate of information-driven decision making on both the clinical and business sides of the hospital. The original systems that Nesson and Glaser had established also incorporated a reporting tool called EX, and a data warehouse called CHASE (Computerized Hospital Analysis System for Efficiency). The analyses and data from these systems formed the core of the Brigham’s balanced scorecard.

Before an effective scorecard could be developed, the Brigham had to undertake considerable work on data definitions and management. One analysis discovered, for example, that there were five different definitions of the length of a patient stay circulating in 11 different reports. The chief medical officer at the time, Dr. Andy Whittemore, and the CCE’s Dr. Gustafson, a surgeon who had just taken on quality measurement issues at the Brigham, addressed these data issues with a senior executive steering committee and decided to present the data in an easy-to-digest scorecard.

Under the ongoing management of the CCE, the scorecard contains a variety of financial, operational, and clinical metrics from across the hospital. The

choice of metrics is driven by a “strategy map”7 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch18#ch18end07) specifying the relationships among key variables that drive the performance of the hospital (see Figure 18.1 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch18#ch18fig01) ). Unlike most corporate strategy maps, financial performance variables are at the bottom of the map rather than the top. In the scorecard itself, there are more than 50 specific measures in the hospital-wide scorecard, and more detailed scorecards for particular departments, such as Nursing and Surgery. The scorecard has also been extended to Faulkner Hospital, a Partners institution that is managed jointly with the Brigham.

10/12/22, 1:51 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken… 10/45

Figure 18.1 Strategy map for Brigham & Women’s balanced scorecard

Dr. Gary Gottlieb, the Brigham president from 1992 to 2009, was the most aggressive user of the scorecard. He noted:

I review the balanced scorecard on a regular basis, because there is specific data that is of interest to me. There are key metrics I examine for trends and if they develop, then I analyze the data to better understand what is going right or wrong. It is one view, but an important one of our hospital. I can look at the balanced scorecard and get information in another way, from a different perspective than I can when I’m

making rounds on a hospital unit, or sitting in the meeting with chiefs.8 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch18#ch18end08)

Gottlieb left the Brigham CEO role to become the CEO of Partners overall in 2010. One of the primary initiatives in his new Partners role is to expand the degree of common systems throughout Partners, so that there can be common data and analytics throughout the organization. Perhaps one day all of Partners HealthCare System will be managed through one scorecard.

Notes 1 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch18#ch18end01a) . This and other details of the Partners LMR/CPOE systems

are derived from Richard Kesner, “Partners Healthcare System: Transforming Healthcare Services Delivery Through Information Management,” Ivey School of Business Case Study (2009).

2 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch18#ch18end02a) . “CIRD, Clinical Informatics Research & Development,” http://www.partners.org/cird/ (http://www.partners.org/cird/) .

3 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch18#ch18end03a) . PricewaterhouseCoopers, “Partners HealthCare: Using EHR Data for Post-market Surveillance of Drugs” (2009). http://pwchealth.com/cgi-local/hregister.cgi/reg/partners_healthcare_case_study.pdf (http://pwchealth.com/cgi-local/hregister.cgi/reg/partners_healthcare_case_study.pdf) .

4 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch18#ch18end04a) . Healthcare Financial Management Association, “Developing a Meaningful EHR,” http://www.hfma.org/Publications/Leadership-Publication/Archives/Special-Reports/Spring-2010/Developing-a- Meaningful-EHR/ (http://www.hfma.org/Publications/Leadership-Publication/Archives/Special-Reports/Spring-2010/Developing-a-Meaningful-EHR/) , Part 3 of “Leadership Spring-Summer 2010 Report: Collaborating for Results.”

5 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch18#ch18end05a) . The 25 meaningful use criteria are described in “Eligible Provider: ‘Meaningful Use’ Criteria,” by Jack Beaudoin, Healthcare IT News, December 30, 2009, http://www.healthcareitnews.com/news/eligible-provider-meaningful-use-criteria (http://www.healthcareitnews.com/news/eligible-

provider-meaningful-use-criteria) .

6 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch18#ch18end06a) . Robert S. Kaplan and David P. Norton, “The Balanced Scorecard: Measures that Drive Performance,” Harvard Business Review (January – February 1992).

7 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch18#ch18end07a) . Robert S. Kaplan and David P. Norton, “Having Trouble With Your Strategy? Then Map It,” Harvard Business Review (September – October, 2000).

8 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch18#ch18end08a) . Ibid.

10/12/22, 1:51 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken… 11/45

19 Catholic Health Initiatives

Deborah Bulger and Evon Holladay

Healthcare organizations sometimes struggle with managing the volumes of data they produce—from financial, clinical, operational systems, and processes. Yet the ability to manage data and transform it into meaningful information yields significant returns to an organization’s business performance. A recent report by Yonek et al. on the characteristics of high-performing healthcare organizations cited several best practices including:

• Establish a systemwide strategic plan with measurable goals and track progress toward achieving them with system performance dashboards.

• Create alignment across the health system with goals and incentives.

• Leverage data and measurement across the organization by, among other things, frequently sharing dashboards and national benchmarks with

hospital leaders and staff to identify areas in need of improvement and taking immediate actions to get back on track.1

(http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch19#ch19end01)

To be effective, information must reach the people charged with improving performance, and must reach them in a timely and appropriate fashion. At the enterprise level, a global, measurable strategy will set direction for the organization. At an operational level, measures that support financial and capacity related activities should be readily available. And for measures related to patient care activities, reporting must reach caregivers in real-time. The challenge: How does a healthcare organization translate data into actionable information for every stakeholder across the enterprise?

About the Organization Catholic Health Initiatives (CHI) is a national nonprofit health organization with headquarters in Denver. It is a faith-based system that includes 73 hospitals; more than 400 physician practices; 40 long-term care, assisted- and residential-living facilities; a home health agency; and two community health-services organizations in 19 states. Together, its facilities serve more than 60 rural and urban communities and provided nearly $500 million in community benefit in the 2010 fiscal year, including services for the poor, free clinics, education, and research. With approximately 70,000 employees and annual revenues of more than $8 billion, CHI ranks as the nation’s third-largest Catholic healthcare system. It is ever moving toward its vision of Catholic healthcare as a vibrant ministry, ready to provide compassionate care of the body, mind, and spirit through the twenty-first century and beyond.

Business intelligence (BI) is a relatively new function for CHI. It is responsible for providing a historical, current, and predictive view of business operations through an enterprise data warehouse. CHI’s patient data warehouse provides information for strategic reporting and core measures of regulatory compliance. The department is partnering with leadership across CHI to develop metric standards and define key performance indicators and best practices benchmarks. The goal is to reduce latency in decision making by having information readily available.

Current Situation Because of the breadth and depth of services provided and geography covered, CHI represents a microcosm of the U.S. healthcare delivery system. The distributed nature of the organization, disparity of systems, and the sheer magnitude of data produced across the enterprise all contribute to the complexity of data standardization. For instance, CHI uses multiple vendors across the enterprise for hospital information and acute care billing systems, clinical decision support, compensation, revenue management, enterprise resource planning (ERP), productivity, and so on. In physician practices alone there are 14 different vendor solutions. To achieve an enterprise reporting model, CHI is leveraging commercially available tools for BI, data marts, extract/transform and load (ETL), and enterprise data warehousing.

Like Yonek et al., CHI recognizes a need for a strategic alignment of technology, information, and stewardship if the organization is to move up the analytics maturity curve. In the book Analytics at Work: Smarter Decisions, Better Results, Davenport et al., explains how organizations can use data and

analysis to make better decisions.2 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch19#ch19end02) Aimed at a broad, multidisciplinary audience, it speaks to employees across their organizations who want to know where they stand now and what they need to do to become more analytical over

time. The DELTA3 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch19#ch19end03) model outlines five key components for deploying and succeeding with analytical initiatives:

• D for accessible, high-quality data

• E for an enterprise orientation

• L for analytical leadership

• T for strategic goals or targets

• A for analytical talent

It is through this model that we describe CHI’s journey toward enterprise intelligence.

Data

Data are the foundation for analytics, and CHI recognizes that managing data across the enterprise requires discipline. The organization has identified three critical steps to establishing an enterprise data model.

10/12/22, 1:51 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken… 12/45

Selection of the Standard

At CHI, the lack of consistent definitions has, at times, led to false assumptions about an individual organization’s performance, creating barriers to an enterprise approach. To first agree on relatively simple definitions, such as whether to include day surgery or lab visits in volume measures, makes it easier to tackle more complex definitions such as adjusted patient days. CHI is designing this model operationally through its governance structure (discussed later in this chapter) and has started the standardization process with acute care, to be followed with physician practices and home care.

Of particular importance are the partnerships CHI has developed with its software vendors to standardize naming conventions within their products. This helps to ensure definitions are consistent even outside the enterprise data warehouse.

Implementation

Once standards are determined, the organization needs to ensure that they are implemented. CHI plans to use data governance to understand business requirements, design data definitions, develop and test metrics, and ensure effective implementation.

A change control process that was started in BI will be adopted across all CHI reporting systems. This process begins with a gap analysis to evaluate

new data definitions or changes to calculations and determines when those changes will be activated. To ensure accountability, CHI uses a RACI4

(http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch19#ch19end04) matrix to assign people who are responsible, accountable, consulted, and informed for activities and decisions that impact implementation.

Ongoing Monitoring

It is critical to monitor data continuously to ensure its integrity, identify new measures, and assess the information behaviors of people who gather and use it. Administrative data are highly standardized, but as new metrics demand a higher level of data accuracy, it is necessary to train people to collect data correctly. One example at CHI is the capture of patient race and ethnicity data. These are key data elements used for monitoring underserved populations and mitigating disparities in care, a critical component of health reform. There are national standards for this—no need to recreate the process—but the data are not always captured accurately upon admission. A high percentage of patient records at CHI were listed as “unable to determine.” Rather than take a punitive approach, the organization chose to address the issue as a function of behavior and ongoing process. Once admission personnel understood how these data were used and the importance to patient care, capture rates improved.

Data Management in Action

An example of a decision that impacts implementation is determining whether data must go to the warehouse for further normalization, aggregation, or modeling, or if they can be viewed directly through BI tools in transaction systems or other reporting systems like cost accounting. The decision is based on the frequency of data needs (e.g., real-time patient census supported by a single variable) and the complexity of analysis, such as projected payments requiring statistical models inherent in the data warehouse.

Enterprise

CHI aggregates financial and operational data supplied by each entity. However, it has not historically provided an enterprise reporting methodology with a standard taxonomy for comparing key business practices. As the healthcare delivery model becomes more disparate—acute, ambulatory, home care, long-term care—it creates some interesting challenges for comparisons at an enterprise level. For example, the acceptable operating margin may be very different for long-term care than acute care, so it is important to understand the reason for the difference and recognize the unique contributions of each care setting. CHI’s objective is to view the two holistically, align around a model of shared accountability, and recognize that long- term care is an equal partner in the delivery model. As long-term care centers are added to the organization—40 centers at this writing—they are treated as part of the enterprise rather than independent entities. CHI’s mission is to achieve the best possible care for the community by ensuring appropriate handoffs and measuring effectiveness across the continuum.

Leadership

Becoming accountable for the care of a broad and diverse patient population means that strategic decisions must be made based on reliable data. CHI has structured its information management model to support an enterprise strategy for decision making (see Figure 19.1 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch19#ch19fig01) ).

10/12/22, 1:51 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken… 13/45

Figure 19.1 Simplified graphic of Catholic Health Initiatives data governance data structure

In this model, senior leaders drive a top down directive for corporate alignment of strategic metrics with measurable action plans.

At an operational level, the information management council is a multidisciplinary group of individuals selected by their senior operational leaders, representing all regions of CHI. The objective of this group is to define priority solutions that will provide stakeholders with timely data so they can make better decisions based on information rather than anecdote.

CHI’s data governance structure is composed of functional vice presidents from across the enterprise. Their goal is to create accountability for standardized information that aligns with functional and operational priorities and external best practices. In addition, smaller workgroups aimed at specific business issues are designed to create an environment where participants can build relationships, have difficult conversations, and solve problems.

Target or Goal

Standardized data definitions allow for meaningful comparisons. CHI is large enough to benchmark results both internally and externally. This allows the organization to define current achievement and set aspirational goals. CHI’s enterprise measurement dashboard is completely transparent, allowing each facility to view every other facility’s results. These open comparisons enable organizations to see who is doing better on various metrics and to share the possibilities. It is acceptable for entities to be different as long as they explore those differences and are driven to improve. BI’s goal is to create energy and power by providing information that facilitates learning.

To drive that energy, CHI is shifting the analytics paradigm to include more exploratory analytics that empower local organizations to ask and answer difficult questions about their results. CHI recognizes two important aspects of this applied learning model:

• Access to truly comparative information—Organizations can “drill” to comparisons on the dashboard at the market level and across all facilities with confidence in the results.

• Collaboration—At a functional level, organizations know each other and ask questions to understand the results and create shared practices. This approach creates an internal consultancy of shared practices based on long-term relationships that can be leveraged as needed.

By comparing outside its own organization, each entity can leverage the value of the collective enterprise knowledge.

Analysts

Analytical talent is necessary to “connect the dots” from data source to end results to provide critical insights. Technology enables analysis, but it is

human capital that most benefits organizations that compete based on analytics. Three levels of analysts are described by Davenport et al.5

(http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch19#ch19end05)

• Senior management sets the tone for analytical culture and makes decisions.

• Professional analysts gather and analyze data and report results to decision makers.

• Analytics “amateurs” use the outputs to perform their jobs.

This is an area where CHI, like many healthcare organizations, continues to learn. At this point in the journey, CHI is on the cusp of Stage 3—still dependent on localized analytics but aspiring to become an analytical organization (see Figure 19.2 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch19#ch19fig02) ) with leaders who are setting the analytics tone of the organization. The next steps are to start aligning analysts—both professional and “amateur”—around a common understanding of data and measurement and to “get everyone out of their silos.” As these capabilities mature, analysts will help build a framework of trust in both the data and shared practices that drive improvement.

10/12/22, 1:51 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken… 14/45

Figure 19.2 Five levels of analytics capability

Conclusion The focus on coordination of care as an industry driver, the expansion of the care delivery model outside acute care, and pivotal leadership changes have created the “perfect storm” for CHI. As it moves into the next phase of analytics maturity, CHI will plan for additional milestones:

• Evaluate the concept of incentives tied to improving performance by measuring the “return on information” investment.

• Develop enterprise intelligence solutions that provide data in a more real-time manner. Ideally every stakeholder should have a common set of standardized information at her fingertips.

• Define an analytics roadmap. CHI has engaged consultants to help the organization accelerate deployment of high-value analytics.

• Find those true areas that need to improve by deploying more advanced analytics.

• Leverage meaningful use requirements as a baseline to create the next level of learning.

CHI views this process as a journey that promises to weave information into its business practices and deliver against predictable milestones in the future. The organization has made significant progress in the last 14 months as its culture has shifted to support an enterprise model, but there is still much work to do.

Notes 1 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch19#ch19end01a) . J. Yonek, S. Hines, and M. Joshi, A Guide to Achieving High

Performance in Multi-Hospital Health Systems, Health Research & Educational Trust, March 2010, http://www.commonwealthfund.org/Content/Publications/Fund-Reports/2010/Mar/A-Guide-to-Achieving-High-Performance-in- MultiHospital-Health-Systems.aspx? utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A+TheCommonwealthFund+%28The+Commonwealth+Fund%29 (http://www.commonwealthfund.org/Content/Publications/Fund-Reports/2010/Mar/A-Guide-to-Achieving-High-Performance-in-MultiHospital-Health-

Systems.aspx?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A+TheCommonwealthFund+%28The+Commonwealth+Fund%29) .

2 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch19#ch19end02a) . Thomas H. Davenport, Jeanne G. Harris, and Robert F. Morison, Analytics at Work: Smarter Decisions, Better Results (Harvard Business School Press, 2010).

3 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch19#ch19end03a) . Davenport et al. DELTA is also the Greek letter that signifies “change” in an equation.

4 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch19#ch19end04a) . RACI is an acronym that stands for responsible, accountable, consulted, and informed and deploys a matrix to assign. There are many references to RACI. This is one of them: http://www.projectsmart.co.uk/how-to-do-raci-charting-and-analysis.html (http://www.projectsmart.co.uk/how-to-do-raci-charting-and-

analysis.html) .

5 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch19#ch19end05a) . Thomas Davenport and Jeanne G. Harris, Competing on Analytics: The New Science of Winning (Harvard Business School Press, 2007).

10/12/22, 1:51 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken… 15/45

20 Analytics at the Veterans Health Administration

Thomas H. Davenport

The Veterans Health Administration (VHA), a unit of the U.S. Department of Veteran Affairs (VA), provides medical assistance to military veterans through 152 hospitals and medical centers, 784 outpatient clinics, and more than 100 long-term care facilities such as nursing homes. It serves a veteran population of more than 22 million and is the largest medical system in the United States.

The VA had some historical issues with quality of care for veterans, but for the last decade, it has performed well in that regard. Under its leader Dr. Kenneth Kizer, in the mid-1990s the VHA embarked on a major transformation in care quality and cost reduction. As one aspect of the transformation, the VA shifted resources from inpatient to outpatient care. At the same time, it decreased staffing while improving patient outcomes. For example, in only the four years from 1994 to 1998, the VA made the following changes in care programs:

• Closed 54% of acute care beds.

• Decreased bed-days by 62%.

• Decreased staffing by 11%.

• Increased the number of patients treated by 18%.

• Increased ambulatory visits by 35%.

• Instituted universal primary care.

• Reduced morbidity rates by 30%.

• Reduced mortality rates by 9%.

• Eliminated 72% of all forms.1 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch20#ch20end01)

Veterans routinely rank the VA system as having better quality than other treatment alternatives, according to the American Customer Satisfaction Index. In 2008, the VA had a satisfaction rating of 85 for inpatient treatment, compared with 77 for private hospitals, and VA outpatient care outscored

private hospital outpatient care by three points.2 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch20#ch20end02)

The VA’s EMR System and Related Analytics The VA was one of the first large government care providers—or large providers of any type—to put a standard electronic medical record in place throughout the system. Originally known as the Decentralized Hospital Computer Program (DHCP), it was adopted in the 1980s. DHCP is still the core of the health information system in individual medical centers, though it has gained many new functions over the years. Renamed VistA (Veterans Health Information Systems and Technology Architecture) under Kizer in 1996, it was made available to other healthcare organizations under an open source arrangement. It included functionality such as wireless laptop access in patient rooms, bar coding of medications, electronic signatures for procedures, and access to online images. The VA had also recently added an online patient portal to VistA functionality. The portal reminded patients about allergies and medications, listed past and upcoming visits to VA facilities, and allowed e-mailed communications with care providers. Some patients had automated links between home monitoring devices and their VistA medical records.

VistA and other VA patient data were also increasingly being used for analytical purposes. In VistA itself, clinicians could create, for example, a chart of risk factors and medications to decide treatments. They could also search VistA records to find out, for example, if veterans were showing symptoms of

diseases related to Agent Orange exposure.3 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch20#ch20end03) When a VA hospital in Kansas City noticed an outbreak of a rare form of pneumonia among its patients, a quick search revealed that all the patients had been treated with a contaminated

nasal spray.4 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch20#ch20end04) In another example, VA researchers used VistA data to examine 12,000 medical records to explore treatment variations for diabetes across different VA doctors, hospitals, and clinics, and how patients fared under the different circumstances. The findings were then incorporated into clinical guidelines delivered by the VistA system. In the 1990s, the VA began using VistA data to identify underperforming and particularly successful surgical teams or hospital managers with regard to quality and safety

benchmarks.5 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch20#ch20end05)

In addition, the VA made available a variety of SAS datasets for analysis by researchers. These were extracts from a large patient care data warehouse using VistA data, and they addressed topics such as inpatient care and procedures, and outpatient visits and events. The patient records were normally anonymized.

The VA also maintained the Decision Support System, a managerial cost accounting system based on commercial software. It combined clinical and cost data to allow cost allocation to patient care products and services. Fully implemented by 1999, it allowed the VA to integrate expenses, workload, and patient utilization and outcomes. There were also data warehouses for each regional Veterans Integrated Services Network (VISN), and a Pharmacy Benefits Management Services database of all prescriptions issued by the VA.

To facilitate access to these tools by researchers and analysts, the VA maintained a VA Information Resource Center, an online portal that served as a guide to available research data, tools, and services.

All of these systems and analytical tools are employed throughout the VA. They are evidence that the VA is a leader in both clinical informatics and the performance improvements based on them. Some of the specific groups who use the data and perform the analyses are described in the following sections, along with some of their analytical activities.

10/12/22, 1:51 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken… 16/45

Analytical Groups and Initiatives In addition to the earlier VistA analytics efforts, many analytical initiatives involving patient care at the VA over the past several years have taken place in the office of the Under Secretary for Health for Quality and Safety. That office includes three units that perform analytical work for the VA:

• Quality and Safety Analytics Center (QSAC)—This umbrella unit supports analysis and learning from the vast amount of patient data available from VistA. Not all of the data are useful for analytical purposes, so one of the tasks of QSAC is to determine which data elements are suitable for detailed analysis. QSAC has two specialized units under it: Inpatient Evaluation Center and Office of Productivity, Efficiency, and Staffing.

• Inpatient Evaluation Center (IPEC)—IPEC houses more than 20 quantitative analysts. It focuses on analysis of data on inpatient care to improve patient outcomes. It provides data and analysis to care providers and managers, focusing initially on intensive care units (ICUs), and later on acute care more generally. One of its first activities was to develop risk-adjusted metrics of patient outcomes that could be used throughout the VA. The risk adjustment method was based on data from an extraction program run at each medical center. The data were then analyzed to create reports that compared risk-adjusted mortality and length of stay to the medical center’s adherence to process measures. ICU performance was compared against “average” and “best” performance benchmarks. ICU reports were segmented by specific patient groups such as the type of intensive care unit, severity of illness, or admission diagnosis, or procedure. Using a Web-based database application created and supported by the VA IPEC, hospitals could also track their hospital-acquired infections in the intensive care units.

IPEC also supports the identification of other evidence-based practices to improve the care of veterans and their families. It addresses practices related to central line infections and ventilator-assisted pneumonia, and also focuses on practices that avoid urinary tract infections.

• Office of Productivity, Efficiency, and Staffing (OPES)—The OPES, which includes mathematicians and economists, undertook a variety of projects on the business side of VHA performance. It assesses such topics as clinical productivity, staffing levels, and overall efficiency. One key focus is the productivity of primary care physicians and some specialists, segmenting clinicians by teaching mission, practice setting, and patient complexity. Clinicians as well as VA system leaders are the primary audience for OPES analyses.

Analytics is also performed in other VA groups, although the purpose of the analyses is often research and publication (as it sometimes is in IPEC and OPES) more than changes in medical processes and treatment protocols. For example, the VA’s Health Service Research and Development organization conducts rigorous research and publishes it in medical journals. There is also an outcomes analysis group in the VA’s Surgery department, as well as other quantitative analysts in the Policy and Planning organization and a predictive modeling group in the Patient Treatment File organization.

The VA Office of Information and Technology also maintains a series of data warehouses (both a centralized “corporate” data warehouse and several regional warehouses), and a Business Intelligence Service Line to help with field operations information needs.

Quality Metrics and Dashboards A key component of the 1990s care transformation at the VA involved a new framework by which to measure quality, and holding senior managers accountable for improvements in performance measures. The quality care framework includes morbidity rates, mortality rates, longevity (for example, one-year survival rates), functionality scores, and performance indicators. Throughout the VA, patient function is measured by a version of the SF-12 Health Survey. Performance indicators include a prevention index (for example, vaccination rates, cancer screening) and chronic disease care indices (such as hypertension control).

These metrics proliferated through multiple systems, displays, and dashboards. One of the early activities of the quality and safety analytics groups was to establish a single quality and safety dashboard for the VA. Called Links, it contains both process and outcome measures, with about seven mortality metrics for each facility.

In addition to Links, the VA’s analytics groups also experimented with a variety of other dashboards and displays. Quality information, for example, is presented on statistical process control charts. Different medical facilities are compared on a “variable life adjusted display.” The quality and efficiency levels of different facilities are compared in a “stochastic frontier analysis” displaying efficiency as a frontier of quality. Information in dashboards is often color-coded, and whenever possible the display shows trends and movement.

The quality and safety analytics groups also worked with new metrics, including medical center readmission rates, mental health readmissions, and 12 different ambulatory care conditions. If the 12 conditions are present, patients should not be admitted to medical centers. IPEC also developed measures of patient case severity and tracked whether particular facilities were admitting the types of cases that they were prepared to address.

Many of the metrics were presented using geographical comparisons. The researchers found a high degree of variation in quality and efficiency across the various VA facilities. The metrics and reports were intended to identify underperformers and best practices.

One key challenge at the VA is the amount of data in the organization. Analysts work to provide not just more information, but greater insight. For example, VA analysts generated a comprehensive Brief Analytical Review of quality and safety findings. It includes a variety of data sources, from internal quality data to peer reviews, surveys, and Joint Commission reports. All of the data go into a single report that “tells a story with data.”

The goal of these analytical initiatives, of course, is to stimulate improvement in quality, safety, and efficiency—particularly in problematic facilities. In 2010 the VA began to post many of the measures on the Internet, including various death rates, intravenous line infection rates, ventilator-acquired pneumonia rates, and readmission rates. The VA mounted interventions for hospitals that fall into the bottom decile of national results, and some doctors and administrators can lose their jobs. One article mentioned:

“The VA secretary pays attention to this,” says William E. Duncan, the agency’s associate deputy undersecretary for health quality and safety.

“Unless people in the VA system have an organizational death wish, they will pay attention to this, too.”6

(http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch20#ch20end06)

10/12/22, 1:51 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken… 17/45

These efforts show clear payoffs. Central line infections, for example, were reduced by two-thirds. Similar reductions were achieved in the incidence of ventilator-associated pneumonia.

Possible Future Uses of Analytics at the VA Much of the past work of analytics groups at the VA involved reporting of the organization’s various metrics. However, analysts are beginning to focus on ways to predict and optimize important phenomena in the care of veterans.

For example, in 2011 the analytics groups at the VA, particularly the IPEC, were exploring the use of a neural network to predict the most likely high- risk sites. Four or five percent of facilities fall into that category each quarter, and analysts want to predict and address poor outcomes ahead of the problem. More broadly, the analytics groups at the VA want to get into greater use of predictive modeling and optimization.

In addition to inpatient and outpatient care, the VA’s analytics groups are also beginning to address “fee care,” or care provided by non-VA facilities and clinicians for a fee. The VA spends more than $4 billion each year in external fee-based care. In general, research found that VA care is less expensive than that provided externally for a fee, and the VA analysts wanted to learn who the outliers are in buying fee care, under what circumstances fee care is used, and how effective it is.

Many of the VA’s future analytical activities likely involve analyses involving people and the human skills to do analytical decision making. One approach might involve exploring the relationships between a facility’s employee attitudes and leadership behaviors, and health outcomes at those facilities. There is already some evidence that facilities with poor leadership scores and low psychological safety levels among employees have poor health outcomes. In the near future, the VA’s analytical leaders hope to convert these research findings into intervention strategies.

There might also be future initiatives to nurture the analytical skills of VHA managers and clinicians. The quality and safety organization established an Analytics Academy, which offers quarterly training sessions around the country. Average attendance has grown from 25 to 80 per class, and these efforts might expand in the future. The new analytics organization also promises greater collaboration between analytics providers within the VHA on the key problems and decisions of the organization.

While the VA has made considerable strides with analytics and is clearly among the most aggressive users of clinical analytics in the United States, there is no complacency among analytical leaders. Pockets of the organization still resist evidence-based change, and analytical executives discuss helping the VHA overcome the “five stages of data grief.” The goal, of course, is to get past any grief about poor performance and fix the problem. This attitude continues to serve the VA well as it moves toward a more analytical future.

Notes 1 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch20#ch20end01a) . Speech by Kenneth Kizer, “Reinventing Government-

Provided Healthcare: The ‘New VA,’” Leonard Davis Institute of Health Economics, University of Pennsylvania, 30 April 1999.

2 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch20#ch20end02a) . “Vets Loving Socialized Medicine Show Government Offers Savings,” Bloomberg, October 2, 2009.

3 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch20#ch20end03a) . Gary Hicks, “A. O. Miner: Speeding Benefits to Vietnam Vets,” Vanguard, U.S. Dept. of Veterans Affairs, Nov./Dec. 2010, p. 6.

4 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch20#ch20end04a) . Phillip Longman, “The Best Care Anywhere,” Washington Monthly, Jan./Feb. 2005, online at http://www.washingtonmonthly.com/features/2005/0501.longman.html (http://www.washingtonmonthly.com/features/2005/0501.longman.html) .

5 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch20#ch20end05a) . Phillip Longman, “Code Red,” Washington Monthly, July- August 2009, online at http://www.washingtonmonthly.com/features/2009/0907.longman.html (http://www.washingtonmonthly.com/features/2009/0907.longman.html) .

6 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch20#ch20end06a) . William M. Burton, “Data Spur Changes in VA Care,” The Wall Street Journal, March 29, 2011.

10/12/22, 1:51 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken… 18/45

21 The Health Service Data Warehouse Project at the Air Force Medical Service (AFMS)

Albert Bonnema and Jesus Zarate

The Air Force Medical Service (AFMS) works in close coordination with the Assistant Secretary of Defense for Health Affairs, the major air command surgeons, the Departments of the Army, Navy, and other government agencies to deliver medical services for more than two million eligible beneficiaries. Beneficiaries include active duty, family members, and retirees, during both peacetime and wartime.

The AFMS consists of approximately 38,000 officers, enlisted and civilian personnel, plus an additional 15,000 members assigned to the Air Force Reserves and the Air National Guard. The AFMS has an annual budget of approximately $5.4 billion and runs 75 military treatment facilities, including 16 hospitals and medical centers.

Vision and Mission The AFMS’s vision is to provide quality, world-class healthcare and health service support to eligible beneficiaries anywhere in the world at any time. The AFMS’s mission is to provide seamless health service support to the United States Air Force (USAF) and combatant commanders. The AFMS assists in sustaining the performance, health, and fitness of every airman. It promotes and advocates for optimizing human performance for the war fighters, including the optimal integration of human capabilities with systems.

The AFMS operates and manages a worldwide healthcare system capable of responding to a full spectrum of health requirements. This ranges from providing care in forward-deployed locations to comprehensive preventive care.

The AFMS Office of the CIO Under the direction of the AFMS Surgeon General, the AFMS Office of the CIO (OCIO) manages the strategic vision, implementation, delivery, and interdependence of all AFMS information management (IM) and information technology (IT) programs, including clinical information systems and healthcare informatics. Informatics responsibilities include portfolio and program management, budgeting, and stakeholder leadership, as well as oversight of project execution.

The OCIO addresses daily tactical challenges born of multiple concurrent projects to modernize the AFMS’s information management and information technology. The OCIO is focused on creating next-generation capabilities and infrastructure while ensuring business continuity.

Efforts are currently under way to reshape the OCIO, bringing in new skill sets and creating an infrastructure and architecture that can support the AFMS for the long term.

AFMS’s Modernization Challenges The AFMS already meets many “meaningful use” mandates and has done so for years. But to operate more efficiently and cost effectively, the AFMS’s IM/IT infrastructure must become modernized and integrated. Among AFMS’s key challenges are

• Data integration—The AFMS has had electronic capabilities since the 1990s and has used electronic health records (EHRs) since the early 2000s. The amount of data that the AFMS has is astounding. Currently the AFMS receives 400 different data feeds and near real- time data from 101 sites around the world. These data feeds include roughly eight million transactions each day.

However, data acquisition and integration has been developed organically and on a solution-by-solution basis, without alignment to any common standards or platforms. Data has been siloed in a wide range of legacy systems, some of which are difficult or impossible to support due to a lack of resources, documentation, or skill sets.

The lack of data centralization and data integration limits the value of the data and creates significant costs to maintain the databases and legacy systems. A few years ago, the AFMS decided that it had to modernize its data infrastructure to centralize and integrate its data.

• Information deliverables—Users of AFMS data sometimes feel they lack access to important data or must spend a significant amount of time finding, acquiring, improving, and personally integrating data to build the information artifacts they need. Not all information is delivered in a user-friendly way, it may take too long to access, and it is often designed just for a single-solution purpose. In addition, there is a lack of capacity (both technological and resource capacity) to support what customers want or need. As a result, many key consumers must be turned away to seek out other solutions.

• Economies of scale—Current tools lack the capabilities that are needed to quickly answer queries and create presentation-quality deliverables. The complexity of data has long and steep learning curves. The variety of technologies employed in data integration prevents economies of scale and hinders the development and institutionalization of standards and best practices.

• Lag times from idea conception to realization—Currently the total lag time from an idea for an analytical undertaking until a final output can take three to five years, sometimes longer. This long lag occurs as data has to be aggregated, resources are allocated, development and testing occur, information assurance (the government’s term for data security) is performed, and implementation takes place.

• Creating a skilled workforce—A key part of modernization is growing a next-generation workforce that has the technical skills to use the data that is available.

• Strategic alignment—The OCIO faces the challenge of ensuring that all vendors, contractors, and key AF resources are aligned to a consistent, clear, and widely communicated IM/IT strategy and are empowered to succeed in the missions the AFMS has tasked them with.

10/12/22, 1:51 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken… 19/45

Analytics at the AFMS As with many large organizations, the ability to leverage data from disparate systems to provide usable information for operational and analytical reporting is an ever-growing challenge. Along with the general data explosion that has occurred, the healthcare industry has the unique challenge of standardizing data across health systems to enhance “point-of- care” delivery and use data for research analytics.

For over a decade, the AFMS has provided business intelligence (BI) services via two separate offices: the Health Informatics Division (HID) and the Health Informatics Suite (HIS). By providing registries and action lists, the HID enables clinicians to manage their complex and chronic patient population for effective disease management. The HIS develops solutions to assist the Healthcare Integrator (HCI) with initiatives such as provider schedule management, cost of care, and business planning.

The Health Service Data Warehouse Project The Health Service Data Warehouse Project (HSDW) was driven by the challenges and pain points described previously. These challenges dictated the need for this data warehouse project to follow the approach of data infrastructure first and information delivery second, which focuses data integration resources on modernizing the HID’s data acquisition and processing functions.

This “urban renewal” effort enhanced the HID’s capabilities with the implementation of best-of-breed data integration software and more robust data architecture and infrastructure designed to scale for future growth and needs.

Previous attempts to modernize the AFMS’s information assets focused heavily on consolidation and virtualization without focusing on data integration. While the colocation of database assets is important, it’s equally (if not more) important to devote the time and resources to truly integrate the data. Without focused modeling and integration of the data, the organization will simply have all of its redundant information assets in one place without realizing economies of scale and the benefits of having one source of information.

The most critical factors for this project were

• Data acquisition and transformation—Previous data-acquisition processes used disparate technologies, were often antiquated, performed poorly, and were undocumented. Key knowledge workers to support the code have departed the HID, and the mix of technology skills required to sustain the current operations is too varied, increasing resource cost and hindering the creation of common technology standards. The solution involves transformation of the enterprise’s data-acquisition processes to a centralized, completely integrated data warehouse.

• Management—Historically, business rules, metadata, and system documentation have not been centrally managed. HID capacity issues and varied technology have retarded the documentation process. Change management has frequently not been formalized, creating a moving target when process remediation projects are undertaken. And formal service level agreements for batch windows, performance, and system/data availability have not existed. In creating the HSDW, managerial processes have been revised to address each of these issues.

• Information delivery—Delivery of information has consisted largely of relatively static “push reports” or creations from analysts derived from hands-on data scripting and SQL Queries. Self-service BI has not existed. Along with the HSDW, the AFMS is creating specific data marts for various purposes (like the patient-centered medical home) and has created a portfolio of dashboards.

Implementation

A key to the success of the HSDW was having a champion in the organization who articulated a vision for BI and evangelized that vision. It shows the importance of having strong leadership support.

The HSDW implementation process took 12 months and consisted of the following steps:

• Requirements

• Design

• Development

• Test and Configuration

• Deployment

• Sustainment

Any data warehousing or BI project requires the right talent throughout the project life cycle. The key role facilitating the disparate groups involved in implementation is the analysts who bridge the invested parties: business, clinical, and technical. The many roles involved in the implementation process include a DW/BI lead; a work streams and requirements coordinator; lead and senior information architects; data modelers and analysts; SAS BI architects and developers; an SAS Center of Excellence lead and administrator; code and GUI developers; extract, transform, and load (ETL) architects, admins, and developers; and training, metadata documentation, and subject matter experts.

Results and Benefits

From a technology perspective, the architecture that has been developed is flexible enough to support both simple queries and complex analytics. The data now available can be accessed in near realtime. Users can analyze summary data and granular details.

The HSDW acquires, integrates, and stores the data once so that they can be repurposed. The logical components of a mature data architecture that support enterprise data warehousing and business intelligence include

• Integrated HSDW—Data are modeled and related according to business process and workflow.

10/12/22, 1:52 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken… 20/45

• Data marts—Data are contextualized and accessible in a user-friendly form.

• Operational data store (ODS)—Data are persisted in near real-time for operational needs. This has the added benefit of streaming, which limits periodic batch extractions of significant size.

• Data presentation/reporting—There is a first tier of canned reports, charts, and tables to support simple users with frequent and recurring artifacts (SAS Enterprise Business Intelligence (EBI)).

• Advanced analytics—This includes a multitier BI tool capable of servicing complex queries and ad hoc data exploration and analysis (SAS EBI).

In addition to these technical benefits, having integrated data will enable clinical improvement, increase the satisfaction of users, and lower costs by reducing manual support time for disparate databases. It will also lead to greater efficiency through automation and will serve as a valuable asset for clinical research.

Lessons Learned

• In the AFMS, convincing all of the stakeholders to give up their information assets was a two-year process that eventually required an executive directive. An organization cannot underestimate the sensitivity associated with giving up control of data.

• Tool and technology selection becomes much easier when you realize that the talent acquisition to use the tool and/or technology is much more difficult.

• “Cyber warfare” will be a major technological trend to overcome.

• You can never do enough project planning, but too much planning is a threat to stakeholder perseverance, especially when they just gave up their information.

• The biggest mistake in planning occurred during the deployment of the HSDW’s historical data load. This process exposed critical gaps within the existing architecture and infrastructure in the areas of role augmentation, storage capacity, and performance.

Next Steps

The focus of the HSDW on data acquisition, integration, and storage is critically important. However, to realize the full potential of data integration, the AFMS is focusing on data presentation, visualization, and delivery. The following actions are under way to deploy the “next level” BI to the AFMS:

• Using Informatics.

• Developing and delivering a platform for measuring clinical quality.

• Standing up an analyst-friendly SAS capability that allows analysts to focus on analysis rather than coding or data acquisition/integration.

• Developing a baseline for analyzing and delivering meaningful clinical research on de-identified data sets. Also, developing plans to support public/private access to data for research purposes.

• Providing intuitive user-friendly access to providers and managers to measure and improve their own performance.

• Creating reusable, repeatable processes and best practices around BI, and creating a BI Center of Excellence for sharing and developing AFMS- focused methodologies.

• Developing infrastructure and adopting new technologies to meet growing data, data warehouse, and BI needs.

• Integrating additional data into the HSDW.

10/12/22, 1:52 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken… 21/45

22 Developing Enterprise Analytics at HealthEast Care System

Thomas Davenport

HealthEast Care System, an integrated provider network based in St. Paul, Minnesota, is the largest provider of healthcare services to the eastern metropolitan area of the Twin Cities. Consisting of three short-term, general acute care hospitals and one long-term acute care facility, it was created in a 1986 merger of several faith-based hospitals and home care organizations. In 2011 HealthEast had 7,000 employees and 1,400 physicians on staff.

In 2005, HealthEast embarked on a multiyear plan to become the benchmark for quality care in the Twin Cities. “The HealthEast Quality Journey,” as the institution referred to its plan, was focused on making improvements in a variety of industry-standard clinical quality metrics, as well as internal metrics of process, operational, and workforce excellence. The HealthEast Quality Institute (directed by Dr. Craig Svendsen, Vice President and Chief Medical Quality Officer) was responsible for establishing goals and metrics. The institution’s Informatics Department (directed by Dr. Brian Patty, Vice President and Chief Medical Informatics Officer) worked on incorporating improvements into everyday practice through the use of clinical information systems. The Medical Executive Committee addressed the topic of physician engagement with quality measures and care processes.

By 2010 and 2011 these steps had begun to result in substantial quality improvements. On almost all metrics, HealthEast had shown distinct improvement, and the provider led others in the market on key quality and patient satisfaction criteria. HealthEast focused particular attention on specific medical problems, such as ventilator-associated pneumonia (VAP). After implementing a set of process metrics and related order sets (the “VAP bundle”), VAP incidence improved dramatically. There were no incidents of VAP in 2010 in any HealthEast hospital.

In 2010, Thomson Reuters ranked HealthEast one of the top ten U.S. health systems based on a collection of clinical performance and patient care metrics. McKesson, a provider of information systems to HealthEast, gave the organization one of only two Distinguished Achievement Awards in 2010. According to the text of the award:

Two years ago, HealthEast created a centralized command center using electronic tracking boards to help monitor patient flow in real-time 24/7 throughout the HealthEast hospitals. As a result, in less than a year, patient waiting times dropped in the emergency departments,

patient satisfaction scores jumped 36 percent, and ambulance diversion hours decreased 63 percent.1

(http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch22#ch22end01)

The HealthEast leadership was proud of its quality improvement achievements, but felt that there were additional areas to address if the organization were to continue its upward trajectory in care quality and patient satisfaction. One key area to address was enterprise analytics—the analysis and reporting of data across the entire enterprise, with a focus on prediction and not just reporting. It would also be increasingly important in the near future to integrate clinical, operational, and financial information. The pressure on U.S. providers to become accountable care organizations (ACOs) meant that clinical decision support and financial decision support would both need to influence patient care decisions. These capabilities existed independently at HealthEast and were difficult to integrate. Therefore, executives at HealthEast had been discussing the need to create an enterprise analytics capability.

Assessing and Integrating Enterprise Analytics Capabilities The Informatics Department at HealthEast had begun to assess the organization’s analytical capabilities as early as 2008. The department’s leader, Dr. Brian Patty, believed that analytics were critical to HealthEast’s continuing quality journey. He asked Skip Valusek, an industrial and systems engineer with considerable experience in process improvement and analytics, to assess analytical capabilities across the organization. Valusek conducted a survey of IT and managerial employees, and found that on a five-point scale of analytical capabilities, most respondents thought HealthEast was in the

middle at Stage 3.2 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch22#ch22end02) The surveyed group felt the organization had strengths in management sponsorship for analytics, analytical culture, and the analytical skills of HealthEast staff. The greatest weaknesses were judged to be in the data and information technology support for analytics. Valusek conducted another survey in 2010 and found similar results.

Patty and Valusek presented the results of the surveys at a regular monthly meeting of HealthEast’s senior management team in early 2010. There seemed to be widespread agreement that the issue was important. Comments at the meeting included, “Enterprise analytics should be a component of our strategy,” and “This is critical.” In terms of implementation, someone pointed out that, “This requires a vision and steps.”

The ownership of the enterprise analytics issue, however, was not firmly established at the meeting. Dr. Patty had initiated the discussion, but his organization was busy finishing the implementation of an electronic health record (EHR) for HealthEast. After the management meeting, Patty met informally with several senior executives who might have some interest in owning and managing enterprise analytics. None seemed to want to own the function. Patty concluded that it should be housed within his Informatics Department but that the establishment of an enterprisewide analytics organization would have to wait until the EHR had been fully implemented.

Designing the Enterprise Analytics Organization By mid-2011, Dr. Patty felt the time was right to design and implement the new analytics organization. The EHR project was nearing completion, and in another executive session in July 2011 the executive team reiterated its support for enterprise analytics. However, the climate was somewhat less receptive for creating a new organization. Because of continuing pressure on reimbursements and the need for greater efficiencies in its care of patients, HealthEast needed to cut $50 million from its 2012 budget. A substantial number of new hires in analytics would be difficult to justify.

Therefore, Dr. Patty planned that most of the analytics staff would transfer into the department from other parts of the organization. This was feasible given that there were pockets of analytical expertise all around the organization. He envisioned three teams within the analytics organization:

• Enterprise data team—This team would focus on development and maintenance of a new enterprise data warehouse and data sourcing activities to yield “one version of the truth.” It would include database administrators, ETL (extract, transform, and load) staff, and data

10/12/22, 1:52 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken… 22/45

architects. Some finance people who built the mature financial warehouse would transfer into this group. The enterprise data warehouse would eventually have data coming not only from the HealthEast hospitals, but also from clinics, home care, external EHRs, and physicians’ offices.

• Analytics team—This team would use the data from the warehouse to do their analyses. Some of the people for it, including its proposed leader, Skip Valusek, would come from the informatics department. Others would come from existing groups within the organization where some analytics were already being performed. The Analytics team would not only generate analytics, but would also help with interpretation and process improvement based on the results of analyses.

• Reporting team—This team would develop a strong and largely automated reporting infrastructure, ending the current fragmented reporting and manual “data cobbling” practices. Most of the personnel for this team would come from the Quality Institute and the informatics department. HealthEast used a tool called MIDAS+ for much of its quality reporting, which worked well except for the fact that there were often multiple sources of truth, even for Center for Medicare and Medicaid Services (CMS) reporting. Sometimes reports were also drawn from the wrong fields in the EHR. Automating these reporting processes would be critical for consistency and efficiency of data reporting.

Team members would have their primary affiliations to informatics, but would have “dotted line” reporting relationships to the departments they primarily supported, such as finance and quality.

Patty was anxious to get approval of the new organization and move forward with a higher level of analytical activity. He was particularly focused on connecting data across the continuum of care—sourcing, integrating, and analyzing data across the continuum of care. In terms of predictive analytics, he wanted to focus particularly on predictive models of readmission, working with HealthEast’s case management function. And in terms of the organization’s movement toward an ACO, he wanted to be able to report on clinical, financial, and operational metrics across the continuum of care for individual patients.

Notes 1 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch22#ch22end01a) . “HealthEast Honored for Improving Patient Care with

Information Technology,” press release, August 17, 2010, available at http://www.healtheast.org/press-releases/1107-healtheast-honored- for-improving-patient-care-with-information-technology.html (http://www.healtheast.org/press-releases/1107-healtheast-honored-for-

improving-patient-care-with-information-technology.html) .

2 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch22#ch22end02a) . HealthEast used a model derived from Thomas H. Davenport and Jeanne Harris, Competing on Analytics: The New Science of Winning (Harvard Business Press, 2007).

10/12/22, 1:52 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken… 23/45

23 Aetna

Kyle Cheek

Healthcare payer organizations are undergoing a fundamental change to their traditional business model. Largely in response to cost pressures, payers are evolving from their traditional role as transaction-focused claims processing organizations to producers of analytically derived, information- centric consumer products. Payers have a long tradition of reliance on analytics tracing to their actuarial roots, but the recent transformation transcends the use of actuarial analytics to set rates and reserves. Rather, the move toward greater exploitation of value-laden information is centered on a deeper understanding of healthcare consumer behavior with the intent to influence behaviors and ultimately drive meaningful reductions in healthcare costs. A critical driver behind the heightened expectations for payers to provide information-product leadership is the simple fact that they are the primary consolidator of the largest volumes of the resource most critical to that enterprise—healthcare data.

While efforts to effectively harness the value latent in their information assets have met with mixed results, a handful of payer organizations have emerged as leaders in their analytic endeavors. Aetna is a leading example among the few payer organizations that can fairly be considered bellwethers in that space. Combining a deep analytical competency with mature data management practices, Aetna has positioned itself as a leader among healthcare payers and is able to leverage its analytical maturity for true competitive advantage. The origins of that analytical maturity in Aetna are unique, too, in that they have largely been developed organically. The following brief survey of Aetna’s analytics experience describes its origins,

current organizational structure, and lessons for other payers that aspire to greater analytical maturity.1

(http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch23#ch23end01)

History Aetna’s industry-leading analytics organization, Aetna Integrated Informatics, is owed largely to the acquisition of a deep analytical competency in another organization. Specifically, Aetna’s current analytics organization traces to its acquisition of U.S. Healthcare, Inc. in 1996—an acquisition that included a subsidiary, U.S. Quality Algorithms (USQA). USQA was established by U.S. Healthcare in 1990 to provide insight into the billing and clinical

practices of its contracting doctors and hospitals and identify opportunities to control medical costs.2

(http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch23#ch23end02) Thus, with its acquisition, USQA provided to Aetna a mature healthcare analytics capability as well as a mature analytical staffing model. In addition, USQA provided an established data warehousing practice that provided infrastructure both for the acquired competency and on which Aetna’s future analytical practice could be expanded.

Aetna’s commitment to develop an analytical competency (including the competency acquired with USQA) coupled with a data-driven executive culture combined to position Aetna for a deep reliance on analytics. Oral histories of Aetna’s informatics organizations commonly refer to the primacy of data-driven and metric-centric decision making as a key enabler of the success that has been realized through its unique set of capabilities. So deeply rooted is Aetna’s reliance on this analytical capability today that it is generally viewed as a competitive differentiator and a “showcase competency.”

Organization Since its acquisition, the USQA analytics operation has evolved organizationally into today’s Aetna Integrated Informatics. The responsibilities owned by Aetna Integrated Informatics have grown from the provider evaluation analytics that were the genesis of USQA and now provide analytic support across the enterprise. Reflecting its responsibility for analytic support across functions, Aetna Integrated Informatics is situated within Aetna’s Innovation, Technology, and Services reporting structure and, reflecting Aetna’s positioning of analytics as a critical component in its efforts to improve health outcomes and reduce medical costs, it also shares a dotted line reporting relationship to Aetna’s chief medical officer.

Aetna Integrated Informatics’ expanded scope of responsibilities includes five primary services for its internal and external constituencies:

• Provider analytics—This is comprised of outcome and cost analytics that are used to identify opportunities for outcome and cost improvements among physicians and hospitals.

• Plan-sponsor reporting—This constitutes the regular reporting on cost and utilization trends to covered groups (i.e., employer groups).

• Program evaluation and research—This consists of analyses that are routinely performed to assess the ongoing effectiveness of care management programs.

• Custom informatics—This is comprised of special projects.

• Data warehousing—This category refers to business ownership of Aetna’s data warehouse.

Other core business functions that are aligned with the Integrated Informatics functions, especially through the enterprise data warehouse, are actuarial, underwriting, and marketing. Aetna’s antifraud analytics are maintained apart from the analytical operations in Integrated Informatics.

Two critical features are apparent in the organizational alignment of Aetna’s analytical operations. The first is that the core analytical competency (Aetna Integrated Informatics) is situated as a service organization, positioning it to provide broad support across the organization. Rather than virtualizing the organization, critical analytical functions have been consolidated into a single business unit—allowing greater coordination of analytical priorities under common leadership. Second is its business ownership of the data warehouse asset, which provides fundamental support for Aetna Integrated Informatics and also provides a collaboration point around which other enterprise analytic needs can be coordinated with the central competency.

10/12/22, 1:52 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken… 24/45

Analytics Maturity Model The Davenport-Harris Analytics Maturity Model articulates five levels of analytical competency to describe organizations’ capabilities in this domain. From highest to lowest, the levels of analytical capability are

• Stage 5—Analytical Competitors

• Stage 4—Analytical Companies

• Stage 3—Analytical Aspirations

• Stage 2—Localized Analytics

• Stage 1—Analytically Impaired

Aetna clearly meets the criteria for Stage 4 competency and satisfies the Stage 5 criteria in several areas.3

(http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch23#ch23end03) Perhaps most importantly, Aetna has identified those analytics that are most critical to competitive advantage in a comprehensive analytical strategy. Additionally Aetna has consolidated responsibility organizationally for those priorities that provide a consolidated and aligned view of resource needs (human and technical) and adds the benefit of a largely consolidated enterprise analytical toolkit. Importantly, Aetna also has a strong competency in the critical data warehousing infrastructure and business ownership that is required to drive large-scale analytical efforts. The value of strong leadership around data management requirements by the analytics organization is apparent in Aetna’s planned migration of its warehouse from a plan- and product-centric model to a member-centric model. That effort is positioned for greater success by virtue of the involvement of a critical stakeholder. Similarly, the analytics organization is positioned to better evolve its analytical agenda by virtue of its positioning vis-à-vis requirements for the critical data infrastructure.

Isolated examples do remain of analytics that are not fully consolidated into the centralized informatics operation, namely actuarial, underwriting, marketing, and antifraud analytics. However, most of those examples are aligned with the centralized analytics organization through its maintenance of data warehouse content requirements—a basis for alignment that ensures that those stand-alone analytic areas share a common understanding of analytic priorities and resource requirements.

Aetna’s placement near the top of the maturity model makes it a clear leader among healthcare payer organizations, especially considering that most payer organizations are best classified at Stage 2 on the Davenport-Harris maturity model. Unlike Aetna, most payers have some localized capabilities but do not have a holistic analytics strategy. Most payers also lag Aetna’s maturity in the data warehousing space and do not have an enterprise analytics toolkit. Given the context of most organizations in the payer space, Aetna’s maturity is all the more apparent, as is its competitive differentiation with its deep analytical capabilities.

Bellwether Lessons The most apparent lesson to be learned from Aetna’s bellwether position among payer organizations may be more broadly applicable than the conspicuous differential in size and scale between it and other payers that, with few exceptions, operate on a smaller scale in more limited geographies. The first lesson for aspiring analytical payer organizations is the importance of identifying the strategic drivers that offer the most demonstrable value from analytical enhancement. For Aetna, that strategic clarity was to some extent encompassed by the acquisition of the analytical competency—that is, Aetna’s acquired competency already had a focus at the time of acquisition. This established focus in turn provided a foundational basis for Aetna to expand its reliance on analytics to other business functions.

Another important consideration that emerges from Aetna’s case study is the importance of the relationship of the analytics organization to the underlying data infrastructure on which analytics are dependent. For Aetna that is seen specifically in the traditional role its Integrated Informatics organization has played as business owner of the data warehouse. By virtue of its formalized stakeholder role, Integrated Informatics can directly influence decisions that impact the data resources on which it relies. Integrated Informatics’ role as business owner of the enterprise data warehouse also allows a holistic view of evolving analytical priorities and needs, such as the migration of the warehouse to a member-centric view to drive more patient-centric outcome and cost containment analyses.

The importance of organizational placement is also apparent in the Aetna case study. While Aetna’s organizational model surely will not translate directly to every payer, it does underscore the importance of situating a comprehensive analytical competency such that it is able to serve a broad- based constituency within the organization. In the Aetna example this critical consideration around organizational placement intersects with the extent to which the stakeholder role has been formalized around the management of data assets. While Integrated Informatics is positioned to provide analytical support to a broad constituency, it also provides coordination beyond its direct customers by owning responsibility for the data assets on which the entire organization relies.

Finally, the Aetna experience speaks to the importance of developing an internal analytics competency for payer organizations. Aetna capitalized on a unique opportunity that presented itself from an acquisition. But the more important lesson may be found in Aetna’s commitment to internalize the competency it acquired and integrate it across the organization to drive broad-based analytical value. By consolidating its analytical competency in one organization Aetna promotes alignment of priorities and efficiencies in the management of the underlying infrastructure on which analytics are dependent.

Payer organizations that aspire to analytical maturity would be well-served to consider these critical elements behind Aetna’s bellwether use of analytics for competitive advantage.

10/12/22, 1:52 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken… 25/45

Notes 1 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch23#ch23end01a) . Except as noted otherwise, this chapter is based on a

November 23, 2010, telephone interview with Brian Kelly, MD, Head of Informatics and Strategic Alignment at Aetna, and Kathe Fox, PhD, Head of Consultative Informatics at Aetna.

2 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch23#ch23end02a) . “U.S. Healthcare, Inc., History,” http://www.fundinguniverse.com/company-histories/US-HEALTHCARE-INC-Company-History.html (http://www.fundinguniverse.com/company-histories/US-HEALTHCARE-INC-Company-History.html) .

3 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch23#ch23end03a) . International Institute for Analytics (IIA) official benchmarking study not conducted; competency score and conclusion derived from author.

10/12/22, 1:52 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken… 26/45

24 Employee Health and Benefits Management at EMC: An Information Driven Model for Engaged and Accountable Care

David Dimond and Robert Morison

EMC Corporation is the world’s leading developer and provider of information infrastructure technology and solutions that enable organizations of all sizes to transform the way they create value from their information. Headquartered in Hopkinton, Massachusetts, EMC has annual revenue of $16.5B and 45,000 employees around the world.

EMC’s Driving Partnership in Health program has transformed how the company promotes workforce health, how employees consume healthcare services, and how much the company pays for those services.

• Since 2004, EMC’s healthcare costs have been well below the national trend line, amounting to cost avoidance to date of $223M (see Figure 24.1 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch24#ch24fig01) ). The company achieved that dramatic cost containment while expanding services and without shifting cost to employees.

• More than 90% of EMC’s 22,000 U.S. employees are registered on HealthLink, the company’s personal health management portal, and 95% of them have taken a Health Risk Assessment (HRA). An astonishing 94% of covered spouses and domestic partners have done the same.

Figure 24.1 EMC healthcare costs and cost avoidance

• EMC was the first employer in the world to sponsor an electronic and automatically updated personal health record (PHR) for employees who choose to maintain one. More than 40% of employees are doing so on a regular basis.

• The company’s data warehouse of employee health information enables the personalization of employee experience as a health services consumer, the aggregate analysis of healthcare consumption and cost trends, and the identification of opportunities to lower cost by improving workforce health.

• Its efforts have earned EMC recognition and awards from, among others, the Massachusetts Health Council, New England Employee Benefits Council, Mass Technology Leadership Council, and a UnitedHealthcare Apex Award for innovation in the healthcare experience of employees.

• The company does not keep its success a secret, but rather presents regularly at industry conferences and hosts visits from healthcare providers, other employers, and others interested in learning from EMC’s experience. By working closely with the partners in its employee healthcare ecosystem, as well as sharing its experience, the company strives to shape the evolving healthcare landscape.

Vision and Lessons Learned EMC envisions a world where employers can engage patients and providers, enable health awareness and literacy, influence health and lifestyle behaviors, and drive adoption of patient-centric technologies. The results should include healthier and more productive employees, containment of healthcare costs, and societal change around the delivery and consumption of healthcare services. The right marriage of informed and responsible employees with technology-enabled services should improve communication and increase patient involvement in the healthcare process; improve efficiency by eliminating both time delays and duplication of records, tests, and procedures; and improve patient safety by reducing manual error and enabling more sophisticated vetting of diagnoses and prescriptions.

10/12/22, 1:52 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken… 27/45

Why invest actively in a healthy workforce? Because it means greater productivity and lower total cost through reduced claims, reduced absenteeism, enhanced engagement while at work (“presentism”), and lower turnover. As EMC has demonstrated, an employer can bend the healthcare cost curve favorably by influencing utilization, by steering informed employees to make healthy decisions and to choose high-quality and cost-effective options for care. We’ve mapped EMC’s experience in this information-driven program into a preliminary “needscape” as shown in Figure 24.2 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch24#ch24fig02) .

Through the course of developing this program and the culture that surrounds it, EMC has discovered the following:

• Employee engagement—There are relatively simple descriptive analytics that represent the status of a given employee or a group of employees in a panel. With the trend toward accountable care and the need to activate, engage, and perpetually improve the health of chronically ill or at-risk patient panels, trends in this high-value data are featured in management dashboards.

Figure 24.2 Preliminary “needscape” based on EMC’s information-driven program experience

• Program effectiveness—EMC participates in innovative programs to promote wellness and to manage chronic conditions, and shares the results as a means of analytically informing and motivating employees to participate. A striking example is how a panel of hypertensive EMC employees served in a clinical trial that was the genesis for a program founded with Partners Healthcare’s Center for Connected Health called SmartBeat.

• Finance and contracts—These are the key analytical competencies one would expect in a plan sponsor working with its broker to make decisions related to performance improvement, cost containment, and contracting. As a “big data” company, EMC has a unique opportunity to begin working with analytics partners, with de-identified data, in the areas of predictive analytics and even causal modeling. These types of studies will guide critical decisions related to EMC’s plan tiering and serve as precursors to accountable care management methods.

• Utilization and enrollment—The utilization of services, programs, and clinical resources is tracked and trended over time and serves as a critical feedback loop to support decision making and discovery.

• Experience rating and outcomes—These kinds of analytics are common in the music and travel industries, where consumer experience, quality, cost, and productivity (efficacy) rule—and drive adoption and (sometimes disruptive) innovation. As variation in practice decreases with the precision of treatment, people will make decisions about their care based more on such consumer-oriented inputs.

EMC’s level of maturity, when aligned with the Davenport, Harris, and Morison Analytics at Work framework, is summarized in Figure 24.3 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch24#ch24fig03) , which shows the typical types of questions EMC needs to address.

10/12/22, 1:52 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken… 28/45

Figure 24.3 Questions EMC needs to address

Chronology EMC’s approach may seem enlightened and sensible today, especially after the intensive scrutiny of healthcare delivery in the process of enacting national health reform legislation. However, when Delia Vetter, senior director of benefits, and Jack Mollen, EVP of human resources, began exploring new approaches to employee healthcare almost a decade ago, their peers were naysayers. When Jack keynoted a Conference Board meeting in 2003 and talked about EMC’s vision of consumerism in employee healthcare, the feedback said, “It will never work,” and “Employees will revolt over privacy issues.”

In 2002, Delia, Jack, and EMC’s senior management recognized that double-digit annual increases in employee health costs were untenable for both the company and its employees (costs were projected to double in five years), and that past approaches to cost containment hadn’t accomplished much. They needed to do something different to contain cost without simply shifting cost to employees (as most companies have chosen to do).

EMC is self-insured, and it worked through its healthcare plans to get the best available rates, so there was only so much it could do to influence unit costs. Therefore it had to focus on the utilization element of the cost equation: influence utilization by encouraging workforce health, informing employees about healthcare choices and the true cost of services, and focusing on common, chronic, and potentially costly conditions such as hypertension.

The problem was that the company lacked data and analytics about consumption of healthcare services—data was fragmented across 38 health plans across the United States. And no large company had experience to share implementing the approach they had in mind. To understand the challenges and feasibility of new approaches, Delia got involved in healthcare coalitions, business roundtables, and health plan and provider advisory boards— anyplace she could meet and learn from progressive practitioners. In the process, she also found credible experts to invite to EMC to talk with its senior management team.

Meanwhile, her team gathered data from the health plans to draw an initial picture of healthcare consumption at EMC and identify some initial opportunities to improve employee health while containing cost.

The initial strategy was developed with the help of EMC’s insurance broker and advisor, Willis, in 2002. By 2003, they had worked with WebMD to launch the first (mainly informational) version of the HealthLink portal. They also worked with Ingenix to build a data warehouse consolidating claims information from the 38 plans. In 2004, they introduced a more interactive portal, including the basics of the personal health record—services, prescriptions, diagnoses, costs. Delia recalls: “At first we didn’t realize we were assembling a comprehensive and portable PHR. We were just trying to get useful information about consumption and true cost into the hands of employees. Then we realized that we were breaking new and important ground.”

As the data warehouse added information, it enabled EMC to analyze aggregate consumption and cost patterns and find specific opportunities for promoting preventive care and employee health. It also provided the detailed information to enable interested employees to understand true costs and become more effective consumers. Few employees were aware of the cost of services beyond their deductibles and co-pays.

By 2005, HealthLink and the new approach had become institutionalized, the benefits group was working to an annual roadmap of coordinated initiatives, and the program began to receive external recognition. In 2007, HealthLink was opened to family members. In 2008, the PHR was significantly improved to import lab results, enable optional provider access, and enable portability. The Health Risk Assessment was introduced in 2008. Since then, the portal has continued to add capability, with an emphasis on information exchange and integration with providers, as well as functions for specific programs such as remote patient monitoring.

What’s happening today or on the drawing board? Radiology/imaging results, additional lab tests, provider notes, remote monitoring, and biometric readings are being incorporated into the PHR. Virtual clinics staffed by nurse practitioners are being established at major locations, both to save on doctor visits and to encourage employees to get help when they need it. Remote monitoring is expanding to include diabetes patients. The list of available health programs, clinical studies, and online tutorials continues to grow. HealthLink is going mobile with access and applications on personal devices. EMC is working with providers on tele-health initiatives and anticipating the eventual integration of the PHR with industry standard

10/12/22, 1:52 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken… 29/45

electronic health records (as the latter are defined). With the help of its own state-of-the-art products, EMC is enhancing the security of the PHR and all of its healthcare management infrastructure and interfaces.

EMC is planning to roll out an appropriately revised program to employees in Canada and is working on a vision and approach for more actively promoting employee health worldwide. The company continues to participate in health transformation initiatives in Massachusetts and nationally.

Along this journey, there were two keys to success:

• EMC leverages a wide range of partner organizations: Willis, WebMD, Ingenix, major healthcare providers, major health plans, pharmacy managers, medical research institutions, specialized diagnostic and care providers, and business and industry associations. EMC has built an ecosystem of organizations (some in competition with one another) mutually committed to innovate and cooperate in support of EMC’s strategy.

• EMC has stayed true to that strategy and its long-term objectives. The list of initiatives may be lengthy, but they fit together. Nothing is rolled out in isolation. Everything is evaluated and implemented in the context of leveraging technology to improve employee health.

Employee Experience EMC’s goals for its employees center around activation and engagement. That includes taking advantage of available resources, taking responsibility for being an effective consumer of healthcare services, and learning about one’s state of health and how to live a healthy lifestyle.

EMC employees have access to a wide range of self-service tools and resources, support and encouragement in using them, specific incentives, and controls. The introduction to these resources—and to the health management partnership that EMC wants to forge with employees—is a nontraditional health benefits booklet.

The booklet is “nontraditional” because, while it includes the standard enrollment timetable and overview of plans, premium costs, and special programs, the details and comparisons are maintained and accessed online. The booklet instead focuses on EMC’s health philosophy, Driving Partnership in Health: “We believe that being good consumers of healthcare means focusing on healthy lifestyles, patient safety, quality of care, and understanding the impact we, as individuals, have on healthcare costs...choice and responsibility are the foundation.”

The booklet also discusses healthcare reform: how the Patient Protection and Affordable Care Act changes responsibilities of both employers and individuals. It lays out the timeline of healthcare reform provisions, highlights those that EMC has already met (in part due to earlier reform in Massachusetts), discusses implications for employees, and generally educates them on the changing healthcare environment.

The portal to most resources is HealthLink, an online personal health management site for employees and their family members, secure and available 24/7. Central to the health management process are

• Personal health record—Enables employees and adult family members to review clinical data and discuss them with caregivers, in the process avoiding duplicative tests and procedures, minimizing untoward medication interactions and side effects, and generally participating more actively in their treatment. The PHR is automatically populated (via the data warehouse) with medical information including provider, date of service, diagnosis, lab results, prescriptions, and cost of service (both actual and out-of-pocket). The PHR can receive remote patient monitoring information, and it is downloadable and portable.

• Health risk assessment—Survey provides participants with direct feedback on a series of risk factors, including stress, nutrition, weight, blood pressure, blood sugar, cholesterol, exercise, and alcohol, tobacco, and substance use. The HRA yields a total score (on a 100 point scale), a simple means of focusing attention and a benchmark as people try to raise their scores year over year. Linked to the HRA are guidance on behavior change and lowering health risks, as well as the additional costs associated with specific risk factors. It’s not uncommon to hear EMC employees say “What’s your number?” to others who openly discuss having opted into health improvement programs.

HealthLink provides access to a vast array of medical news, information, and recommendations maintained by WebMD—and automatically suggested based on the individual’s health profile. HealthLink also archives EMC’s monthly health seminars. And it provides important functionality such as

• Hospital and physician selection based on condition, procedure, and location, as well as “best doctors” ratings

• Health alerts, including about potential drug interactions

• Automated way to seek a second opinion

• Healthcare cost tracker broken down by category (e.g., physician, pharmacy, lab) and highlighting both employee and EMC costs

EMC employees and families have access, at no added cost, to a variety of specific health management programs and facilities, for example:

• DASH for Health—A dietary program for reducing blood pressure and attaining other health benefits.

• SmartBeat—Remote patient monitoring of blood pressure. EMC participated in the initial clinical trials and helped demonstrate the value of self-management in reducing high blood pressure.

• LiveHealthy—A customized program, including individual coaching and support, for those at increased risk for or currently living with a serious or chronic condition such as asthma, diabetes, or obesity.

• Quit for Life—Smoking cessation program.

EMC sponsors a variety of onsite health management seminars, often supplemented with webinars and support groups. And it offers state-of-the-art fitness facilities at major locations.

The direct incentive for employees and their family members to engage in their health management is a reduction of about 10% in monthly premiums. To qualify for the lower payment in 2011, employees and their spouses or domestic partners must only complete the HRA. To qualify next year, they

10/12/22, 1:52 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken… 30/45

must also complete biometric screening (that automatically populates parts of the HRA) and participate in the “Choose One and Commit” program. The latter simply asks the employee to do something healthy (e.g., participate in a community walk or softball league) or make use of a specific resource (e.g., hospital advisor, mammogram screening, prostate cancer screening, or just maintaining one’s PHR). The “One” activity can be chosen from two dozen options that include the formal programs such as DASH.

Indirect incentives include the encouragement of colleagues and family members. When a spouse or partner completes the HRA and compares results with the employee, it can lead to strong mutual commitment to raise the scores.

Some employees have hesitations and concerns about the health benefits program. Some are simply uncomfortable with their employer’s activist role in what has traditionally been the private matter of maintaining health. This concern has been addressed with a great deal of education, including individual discussions when needed and consistent, persistent messaging about the purpose and terms of the employee-employer partnership.

Others have particular concerns regarding the security and protection of their health information—no surprise since EMC is in the information management business and has experts on the payroll. This is addressed by explaining the safeguards in place and keeping them state-of-the-art using EMC’s own technologies.

Concerns are also mitigated through controls, and the employee is ultimately in control. Employees can opt out of the health benefits program (though already in Massachusetts and nationally in 2014 individuals are required to carry health insurance or pay a penalty). They can pay the full premium and not avail themselves of the HRA, PHR, or other resources. Or they can participate and exercise specific controls over, for example, whether to automatically import claims data to the PHR, whether and when to permit physician access to the PHR, whether to receive health alert messages, and, of course, security and access settings.

As employees gain insight into the costs of their healthcare—and recognize how EMC has maintained coverage and expanded programs without cost shifting—they buy in to the program. One of the surest signs that the partnership is working is that surveyed employees list EMC as their most trusted source of health information.

Partner Perspective What’s it take to partner with EMC in its program to manage employee health and benefits?

• Commitment—Not just to provide excellent standard services, but to work with EMC in meeting its objectives. To determine who was really on board, EMC asked all of its provider and plan partners to rebid for their contracts.

• Collaboration—All the partners, including competing providers and plans, must participate effectively in EMC’s employee health ecosystem. To build this network, EMC convened a Partner Summit of dozens of organizations to share information, align with EMC’s goals, and work together.

Examples of partner experiences are discussed in the following sections.

Advisor

Willis is a leading global insurance broker and consultancy with a history of innovative approaches to risk management. Bill Schlag and a Willis team have been working with EMC for a decade to shape and execute the strategy to engage employees, empower them with information and services, and outperform the healthcare cost trend. In addition to providing ongoing broker services with providers and plans, Willis has advised and participated at every stage of EMC’s journey—from claims analysis to target common and costly conditions, to organizing employee health and claims data into a repository, to launching and improving the portal, to designing and maintaining the dashboard for executive reporting.

Provider

The director of clinical informatics at a major regional healthcare system shared a provider perspective. The organization has an enviable record of quality care and is focused on increasing access to health services to improve prevention and reduce chronic conditions. Thus, its partnership with EMC is rooted in closely aligned goals and mutual ambition to improve. They recognize the limits of what can be done in a doctor’s office (and that many people are too busy to schedule appointments), hence the value of bringing services to patients (e.g., onsite clinics, in-home monitoring) and influencing patient behaviors. Technology is helping with an attitudinal shift: “I can participate in my healthcare anytime.” The HealthLink portal has a gateway to the provider’s patient portal. And personal health records are transferrable (manually today, with a direct interface in the works).

Plan

For a major healthcare plan, the working relationship with EMC is “definitely not standard.” In addition to the basic role of structuring coverages and participating in care management, the organization is challenged to break down barriers to sharing data (including claims and performance metrics), and to support a variety of special initiatives (from the DASH program to onsite immunizations for EMC employees traveling internationally). For example, a specific data exchange problem relates to populating the PHR with lab test results when some (e.g., pregnancy, HIV) are highly confidential and restricted. The onsite immunizations required streamlining a claims process.

For the plan, EMC is a highly respected reference account. The plan also appreciates how EMC is “rallying the provider community” around preventive care and patient participation. The plan has to find the resources to do new things with EMC, and it “can’t work with everyone this way,” but the opportunities to learn and innovate make the relationship very worthwhile.

10/12/22, 1:52 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken… 31/45

Program

DASH for Health (Dietary Approaches to Stop Hypertension) is administered by the Boston University School of Public Health. The original trials of a dietary approach to lowering blood pressure were conducted in Boston and then around the country in the 1990s. Following the diet, which emphasizes fruits and vegetables, can be the equivalent of taking a standard daily dose of hypertension medication. It has also proven to lower cholesterol, improve bone density, and enable people to lose weight and generally feel better. In 2007, EMC employees participated in a research study demonstrating that for people with specific cardiovascular risk, adhering to the DASH diet can lower healthcare costs by an average of $800/year.

Through claims analysis, EMC had determined to focus on hypertension, so the DASH program was a natural fit. The program is administered mainly online, through HealthLink and the DASH site, though DASH staff also provide onsite orientations. Because success with such a diet can depend greatly on having supportive friends, a recent addition has been the opportunity for EMC employees to invite up to three friends to participate with them.

Platform

For eight years, WebMD has partnered with EMC to deploy and enhance the HealthLink portal. HealthLink is a privately branded version of the WebMD portal configured to support the HRA, PHR (portable to WebMD.com), and a variety of self-directed health and lifestyle improvement programs. The portal also incorporates WebMD’s reference information on conditions and treatments, and important functions such as health alerts (e.g., about potential drug interactions).

Working with EMC provides the opportunity to continue to innovate, especially in “keeping the healthy healthy.” Ongoing initiatives include more seamless access to better integrated data, improved self-service and self-reporting by patients, and better analytics to understand the needs of patient cohorts, evaluate the ROI of programs, and improve health management for the entire covered population. WebMD works with other progressive companies, but EMC is the leader in deploying PHRs, developing innovative solutions, and bringing its partners to the table.

Common Threads

Three common threads run through these partners’ experience working with EMC:

• Aligned objectives, including a commitment to quality, cost-effective care, and preventive care

• Willingness to share both information and expertise

• Opportunity—and imperative—to innovate

The Partner Summit was an eye-opening experience for all involved. It brought together providers, plans, pharmacy managers, medical research organizations and their programs, technology providers, specialized clinics and other service providers, and even the providers of disability and life insurance. This range of players had never gotten in the same room, let alone for purposes of brainstorming about the future of employee healthcare and how to make it better. EMC’s attitude was not, “Here’s how we’re going to operate,” but rather, “We’re here to learn from you—what can and should we collectively be doing additionally, differently, and better?” That stance got the players talking, looking at the data, and working together. Erstwhile competitors were co-opted to behave as co-implementers.

Executive Scorecard Not all senior executives think of a healthy workforce and healthcare cost management as factors critical to business success. Given health and healthcare cost trends, they should, so ongoing communication and education are in order. Delia Vetter’s group works with Willis to issue a quarterly scorecard for EMC’s senior management. It includes

• Health plan costs—aggregate, per capita, and trends. These are also broken down by claim type (medical, pharmacy, behavioral health) and health plan.

• How EMC and its employees share those costs.

• Employee and covered member demographics.

• Top ten listings with costs of hospitals, diagnostic categories, pharmaceuticals prescribed, and therapeutic classes.

• Prescription drug breakouts—generic versus branded, retail versus mail order.

• HealthLink portal usage statistics and trends, including aggregate HRA results.

The scorecard also contains a narrative overall assessment plus brief updates on key initiatives, accomplishments, healthcare topics, and other observations. And it incorporates an illustration of EMC’s three to five year healthcare strategy and mention of awards received as an indicator of influence in the marketplace.

The scorecard is four pages long and distributed both electronically and as a laminated document for the convenience of the several executives who like to carry and discuss it regularly. It informs and educates executives and helps align them around employee health strategy and initiatives. It also makes the value on ongoing investment—including in the technologies of healthcare management—very clear.

About the Data Warehouse The EMC data warehouse, from their partner Ingenix, aggregates claims and enrollment data from each of EMC’s vendor partners, including ADP (eligibility), Medco (Rx), the self-insured medical plans, a behavioral health plan, short and long-term disability insurer, and workers’ compensation, plus HRA information from the Health-Link portal. The data are updated monthly.

Ingenix scrubs the data and loads them into a tool called Parallax i that EMC and Willis use to measure, analyze, and evaluate EMC’s benefits programs (see Figure 24.4 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/ch24#ch24fig04) ).

10/12/22, 1:52 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken… 32/45

Figure 24.4 Average costs per employee by benefit type

The data are also grouped into episodes of treatment for analytic purposes. Data feeds go to HealthLink to populate the PHR and for the Health Alerts messaging program, to Medco for management of the prescription drug program and drug interaction analysis, and to EMC’s disease management vendor.

EMC in the Larger Context EMC’s experience and accomplishments offer lessons both for other large corporations and for the healthcare system at large. How EMC uses information and technology to contain cost and promote employee health confirms some of today’s major trends and previews the likely experience of others.

Data is the foundation; trust is the key. With good data, an organization can analyze the patterns of consumption, cost, and medical conditions, then focus on the problem areas. It can isolate cost drivers and manage costs. It can inform and steer employees to be mindful of quality, convenience, and cost—and to make healthier lifestyle choices. And it can continuously measure and analyze what actions are most effective in promoting workforce health and containing healthcare costs. However, the challenges of data acquisition and integration can be herculean. Providers and plans can be reluctant to share cost, price, and quality data, having long treated it as a proprietary asset. And they’ve sought the cover of privacy concerns.

Today, however, we have the means to protect information in the form of advanced and analytically driven adaptive authentication technology, which is part of the next generation PHR platform. EMC knows from the world of financial services that there is a strong correlation between consumer trust and online adoption, so folding in this technology, which is the de facto standard in online banking, just makes sense. A foundation of data, wrapped in employee trust, enables collaboration, information sharing, and discovery that capitalizes on the new transparency.

There are sufficient data to gain experience in the rapidly emerging model of Accountable Care. EMC has implemented many of the principles and practices associated with Accountable Care Organizations (ACOs):

• Analyze and address the needs of specific patient populations segmented by condition or risk factor.

• Focus on preventive and as-soon-as-needed care, thus reducing preventable procedures and hospital admissions and containing cost both short-term and long.

These approaches work for employers and providers, even when the latter are not structured and contracted as ACOs. They simply make sense as the means of simultaneously improving health and containing cost.

Employers should step up, particularly Health Delivery Organizations considering ACOs. Most large corporations and even nonprofit hospitals have been too much on the sidelines of optimized healthcare delivery, offering employees standard (and often shrinking) benefits while absorbing regular cost increases. In the process, some have reneged on obligations to employees and retirees. EMC has demonstrated the value of purposefully participating in and influencing the healthcare delivery process, of being activist on behalf of both employees and shareholders. A key lesson here is that there are no quick fixes, technological or otherwise. Electronic health records (EHRs) and health information exchanges (HIEs) are not going to get all the data in order, and the organization that plays wait-and-see just sees its costs rise. The employer has got to commit and act locally with its employee data, but with an eye toward eventual interface or integration with EHRs and HIEs.

Healthcare is becoming patient-centric and employers will activate and engage their employees with health information centered “medical homes.” Consumers are being empowered with information and choice. There’s no fighting the trend, so employers—indeed, all the players in the healthcare ecosystem—must embrace it. For employers, that means partnering with employees to build their healthcare literacy and encourage them to make

10/12/22, 1:52 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken… 33/45

sensible choices. That also means using technology to understand the employee-consumer’s experience, set new standards for information sharing and transparency, and provide services compatible with how employees live, work, and behave. More of these services are accessible or consumed in the workplace and in the home because that’s both convenient for consumers and cost-effective for providers.

To compete on analytics, the stakeholders across the value chain may be challenged by the influx of analytics-enabled employers with engaged employees who will behave like savvy consumers in an era of choice and responsibility.

10/12/22, 1:52 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken… 34/45

25 Commercial Analytics Relationships and Culture at Merck

Thomas H. Davenport

The Commercial Analytics and Decision Sciences group at Merck is responsible for assisting with advanced analytics for all of the U. S. primary care, hospital/specialty, and vaccine products at Merck. Its focus is sales and marketing analytics, including customer targeting, segmentation, sales force sizing, promotion response modeling, ROI assessments, and other related analyses. The group’s primary mission is to help senior leaders at Merck make better business decisions with regard to multimillion dollar promotional and sales budgets. It has existed at Merck for more than 15 years.

The group consists of more than 25 full-time staff and several other external consultants. The leader of the group has a Ph.D. in Applied Research and Evaluation, and most staff members have advanced degrees in quantitative fields, including operations research, statistics, and economics. Most of the group’s staff came to Merck from analytical roles in other companies spanning numerous industries, including consulting, large pharmaceutical firms, health services research, physician licensure, insurance, and consumer packaged goods. In addition, the group maintains close partnerships with a variety of external providers of data and analytical software and services.

The Commercial Analytics group has been involved in a variety of key decisions at Merck over the last several years. When Merck reengineered its commercial model for U.S. Sales, the group piloted the model before it was adopted, with test and control groups. Other work has quantified the impact and profitability of virtually all major investments in the physician and consumer channels. The group has also created tools for optimizing sales force sizes and structures along with multichannel programs.

Decision Maker Partnerships The Commercial Analytics group maintains a close set of relationships with internal business decision makers. They have positive comments about the group’s role. One executive, responsible for strategy execution, commented:

A lot of times Commercial Analytics team members were my “thought partners” in implementing the new field organization. Working with them was a good way of thinking something through. We used them as sounding boards. They are very solid problem solvers, and play the role of an objective third party.

The same executive said that the Commercial Analytics team was more useful than an external resource that did similar types of work:

Most of the other firms who did these new commercial models used an external consulting firm. We used them for some tasks, but we had our own algorithms developed by Commercial Analytics. They also found ways to optimize and test the pilots. It gave us a better result, as well as more internal buy-in.

The leader of a new business area that worked with Commercial Analytics also had positive comments about the value of the group’s work and their credibility:

Our business area is a pilot program. We want to show that it drives new revenue and provides better customer support. Commercial Analytics is measuring the impact of the pilot program. They set up a rigorous test-and-control approach.... Commercial Analytics is very familiar with the business. They ask what business questions you are trying to answer, and then they identify how to measure them. They will analyze the data to see if they can answer the questions. Their level of objectivity is what you need to have; we need an independent source.... At times in the past, Commercial Analytics had to tell senior management that their project doesn’t have good ROI. They are very credible when they do that. And if they say it works, there won’t be any doubt about it.

A senior executive at Merck with global responsibilities emphasized the value of having Commercial Analytics involved in the entire decision process:

They should always be at the table when we are making an important decision. I remember when we were evaluating the returns on a major promotional campaign a while back. Commercial Analytics was at the table with us throughout the discussion, and would engage with us in debate. Then they would do analysis to answer key questions. Having them be part of the team is a real competitive advantage for us.

Reasons for the Group’s Success There are undoubtedly many reasons why the Commercial Analytics and Decision Sciences group at Merck has been effective. The members of the group certainly have a high level of analytical skills, for example. Another key factor, however, is clearly the culture and relationships orientation in the group. The leadership team of Commercial Analytics emphasizes the key value of the organization:

The umbrella over everything we do is a culture of motivating team members with the prize that we are here to help our clients make better decisions through the use of our analytic insights and tools. Our rules of engagement are to make your internal client understand that you are there to help them make a better business decision.

The cultural orientation begins with clarity about the mission and responsibility of the organization. The group leader notes:

We’re always objective about our findings. In a way we are the “Switzerland” of marketing and sales at Merck, providing a neutral perspective on those decisions. We work for the shareholders.

The group leader gives an example of how the group’s independence affects its work with internal clients:

A lot of times managers will hear that we can do ROI analysis on promotions. So they come to me and ask if we can help them. I say, “We can do that, but let me ask you a question first. We will find that your promotion was very effective, marginally effective, or ineffective. Can you tell me what actions you’ll take in each of those cases?” We document their answers and how the analytics will tie to them.

10/12/22, 1:52 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken… 35/45

The Commercial Analytics leadership team refuses to have the group engaged in a project if there is no clear relationship between an analysis and the decision to be made.

The group’s strategy execution client confirms this approach:

Commercial Analytics always has the question of, “How’s this going to add value to our business?” at the front of their minds. They aren’t chasing stuff as an academic exercise; they do a good job of checking to make sure that what they are about to tackle has business value. They ask, “What are you trying to get at—maybe there is a better way to get there.” I don’t think it bothers anyone when they push back a little— they do it in a nice way.

One of the reasons that the group is able to work successfully with business decision makers is its emphasis on clear and nontechnical communication about its work. Many of the analyses it undertakes are technically complex, but the Commercial Analytics leadership team devotes considerable effort to translating them into straightforward business terms. The leader of the group describes this process:

We work hard at packaging our results in a way that is very intuitive and easy to digest for our business clients. If an analysis is not understandable to our client, then we failed to provide the appropriate graph, chart, or table. We do not avoid complex methods, but we make sure we can explain them. One of our passions is distilling very complex ideas into simple terms so that businesspeople can understand and apply them.

The executives at Merck interviewed confirmed that the communications approaches are succeeding. For example, the executive leading the new business area noted:

Commercial Analytics communicates clearly to businesspeople. They were able to share their methodology with the marketing leaders whose products we are going to be selling. Since those managers are charged with sales force expense, they need to understand and evaluate our pilot.

The strategy execution executive described the communications ability of Commercial Analytics staff in similar terms:

The members of Commercial Analytics didn’t come up through the sales area like I did, but they know they have to translate their findings into something that is “field-friendly.” I know the folks in Commercial Analytics are always thinking about how to do that translation. I have worked with analytical people who are much more academic—it is more effective to work with Commercial Analytics.

Embedding Analyses into Tools One other approach to improving decisions that the Commercial Analytics and Decision Sciences organization takes is to embed results into small software tools for use by marketing and sales managers in the field. The goal is to help field managers make better decisions by providing decision logic and data for the analyses they typically perform.

The group created a “channel choice simulation tool.” It allows the user—typically the planner of a marketing campaign—to simulate the decision of channel selection for a particular product. The user can play with a variety of scenarios while attempting to optimize the returns on investments across channels. The output of the simulation is a probability of achieving a certain ROI level for a particular product.

Perhaps the most focused analysis tool is one for sales force vacancy management. If a sales rep leaves a particular region, should the manager fill the vacancy? This tool provides qualitative and quantitative analysis to inform the vacancy-filling decision. In a sense, it’s a semiautomated checklist of the factors to consider in filling a sales vacancy. A sales manager’s intuitive feeling about the need for a replacement is a key variable in the analysis.

Future Directions for Commercial Analytics and Decision Sciences The leader of Commercial Analytics and the interviewed clients all feel that the group is providing considerable value for Merck. The key question going forward involves the direction for role expansion. Should Commercial Analytics, for example, expand beyond the U.S. market and provide support for global sales and marketing decisions? Business across Merck has become considerably more global through both acquisitions and organic growth, and the non-U.S. businesses need more analytical help with sales and marketing decisions. The downside, however, would be the possibility of providing too little support for important decisions in the United States, which is the largest market at Merck.

Another option for role expansion would involve more “horizontal” collaboration with other analytics groups across Merck. In addition to Commercial Analytics, Merck has strong analytical capabilities in the R&D/clinical area, as well as in health economics and manufacturing. Thus far, the collaborations among these groups have been relatively minimal. Leadership of Commercial Analytics is aware that some other organizations, both within and outside the pharmaceutical industry, are beginning to view analytics as more of an enterprise-level capability. Thus far, however, the specific benefits of greater collaboration are not clear.

Whatever the future roles of the Commercial Analytics and Decision Sciences organization, the values of independence, clear communications, and assistance to business decision makers in multiple forms will continue. These cultural attributes are an important component of the group’s success and have led to a clear competitive advantage for Merck overall and the executives who have taken advantage of the group’s capabilities.

10/12/22, 1:52 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken… 36/45

Conclusion: Healthcare Analytics: The Way Forward

Dwight McNeill

In this last chapter, I look to the future and discuss some pathways to help the field of healthcare analytics become more effective in improving business and clinical outcomes. I start off by describing the present state of analytics as detailed in the book and conclude that there is much to be done to meet the potential for analytics. I describe the huge agenda, burning platform, and opportunity. I concentrate on the data gold rush and technology as the great enabler and conclude with ways to make change happen.

Analytics As We Know It This book provides the most comprehensive review of the current state of the art and science of analytics in healthcare including the following:

• A variety of perspectives on the daunting challenges facing the various sectors of the industry, mostly from healthcare reform and hypercompetitive market pressures.

• An overview of healthcare analytic’s “DNA”: What it is, how it provides value to the business, what are its various forms, what are some examples of its “secret sauce,” and how it deals with the ubiquitous pressure of balancing the use of more and more data with the rising concern about personal privacy.

• A workbench of analytics methods including using electronic health records (EHRs) to provide meaningful results; measuring, monitoring, and improving providers’ adherence to established clinical standards; reducing medical errors through the use of triggers; finding high cost/clinical need people through “hot spotting”; and using emerging approaches to personalized medicine through integration of genomic and other personal data.

• Examples of best practices through eight case studies on the current state of leading analytics. The common characteristics of these high performing companies are the early adoption and use of EHRs, leadership that clearly articulates organizational mission and goals, the use of clinical warehouses to address organizational needs such as research, the application of analytics to improve business and finance functions, and insights into how to organize analytics for optimal results.

The Gap Between the Cup and the Lip By all indicators, however, the promise of analytics far overhangs its performance to date to improve clinical and business outcomes. The opportunity

number is $300 billion in value annually, according to McKinsey and Company,1 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end01)

that is, if “big” data were used creatively and effectively to improve efficiency and quality. That’s about one-tenth of the healthcare economy and equal to about one-third of the waste in the health care system. Analytics can actually do better than that. The potential for analytics to make significant contribution to reduce costs, improve the top line, develop new healthcare innovations, and transform the way business is conducted is underrated.

At the present time, the best of what analytics has to offer is practiced by a small minority of healthcare organizations. Indeed it may be as small as a dozen or two. These organizations are large, very well organized—for example, as integrated delivery systems—and have core strengths in traditional forms of analytics, for example, research. This may be a case of the rich getting richer because they can afford the investments in analytics and can see beyond the horizon. The everyday view for most healthcare organizations is a constant pivot to keep up with compliance and regulatory demands, competitor pressures, and legacy systems both technical and cultural.

The majority of healthcare organizations today are still in analytics version 1.0. Their mantra is “don’t take my spreadsheets away,” and their mindset about analytics is that it is located in the back office somewhere and practiced by statheads who take care of the reporting requirements and do some research types of things and sometimes provide information that is insightful. Many organizations are working hard to build information infrastructure including the EHR, integrate various siloed data warehouses, and upgrade IT capability. This is usually done “over there” in the IT shop, and those in the shop do not want the data “released” until it is housed in the new data warehouse and is clean and perfect. While they are chasing elusive perfection, the clock is ticking on providing value to the business. And there are missed opportunities to demonstrate analytics value with available and relevant data that can be cast into insights for improving the business today. Finally, many entrepreneurial firms are talking about the art of the possible and stressing the opportunities of new analytics innovations to improve the business. Some are building useful solutions.

Overview of the Way Forward Let’s face it. The supply of tools for analytics have been around for quite some time, including deep knowledge on how to collect, integrate, and report data; a large catalog of statistical methods; and a whole field of theory and practice on how to improve decision making. What has changed and may offer a tipping point for analytics in healthcare to take off and provide value are three drivers that change the demand for it:

• The pressures from governments and market reforms to change the financing, payment, and delivery of care, along with the clinical information infrastructure (EHR), such that care is less costly, of better quality, more transparent, and responsive to the needs of the retail customers (real people) it serves

• The explosion of data, a data gold rush, in terms of volume and diversity

• Technology advances that significantly improve the cost, time, and capacity to mine, process, and leverage huge volumes of data as well as inverting how data is collected, shared, delivered, and controlled through mobile and biometric devices and social media

But, there is a nagging issue. Making change happen is really very difficult. It’s not the tools, data, and technology that matter the most. It’s the sociology. John Eisenberg, the esteemed director of the Agency for Healthcare Research and Quality (AHRQ) during its early days, made the point often

10/12/22, 1:52 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken… 37/45

that if we only translated and applied the research we already knew into practice, it would provide more benefit in terms of improvement in patients’ lives than all the research being done to find new causes and treatments.

Indeed the unrelenting focus on technology may be part of the problem. In this era of “bowling alone,” the cryptic and hectic release of ideas on Twitter and Facebook, and the wish for frictionless solutions that often preclude engaging with people, we lose sight of the fact that change happens largely through communications among people. According to Atul Gawande, the wish for “turnkey” technical solutions has led us to a preference for

“instructional videos to teachers, drones to troops, incentives to institutions.”2 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end02)

Understanding the technology of diffusing innovations like analytics is just as important as understanding the skills needed for advances in computing such as NoSQL, ETL, and ODS.

So, those in the profession of analytics need to know the business, how analytics dovetails with it, the data, the various computing and analytics technologies, and how to make change happen. That’s a tall order. But, the old role does not fit anymore. Let’s face it. Most of us practicing analytics are nerds-geeks-statheads who love numbers more than we like people. We like the quietude of the back office much more than the (social) intensity of the front lines. And we are modest and don’t fully understand nor can we express our power to make a very big difference in healthcare. To use a phrase from Pogo, “We have met the enemy and he is us.” We need to change the ways we do our job.

A Huge Agenda and Opportunity U.S. healthcare faces significant challenges and opportunities to improve business and clinical outcomes. Closing the gap on waste and inefficiency can save close to a trillion dollars. If clinical outcomes were improved to match those of the highest performing countries, almost a trillion dollars in value

could be achieved in terms of productive lives.3 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end03) And if customer engagement

were to match that of the highest performing industries, increased market share could amount to nearly $100 million.4

(http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end04)

The time is hot for analytics to demonstrate its value. The environmental drivers from government and the marketplace that demand a retail transformation, delivery system changes, payment changes, and a focus on outcomes are all compelling and toughen a burning platform for organizational change...that relies on information...and is fueled by high octane analytics. In addition to these drivers there are other emerging factors to shape the value proposition of analytics:

• Behavior change—The pathway to demonstrate value depends on changing behavior across stakeholders, for example, doctors to perform to accepted guidelines for care, business leaders to consider and adopt new innovations, and patients to embrace healthy behaviors. Addressing behavior change at the patient/person/member level is probably most important. This is based on the overwhelming evidence that people’s behavior is responsible for good health more than any other determinant and four times more important than healthcare.

• Data democratization—Healthcare has been stubbornly provider-centric. But Web 2.0 and social media, mobile devices, and a powerful patient engagement movement are changing how care can be provided, purchased, evaluated, and modified.

• De-medicalization of healthcare—It is inevitable that cost pressures and the need for better outcomes will lead to new organizational forms for producing health. For example, medical homes are a great advance, but healthy houses that concentrate on chronic diseases and employ low

cost and friendly care by people “just like me” in my home and community are overdue.5

(http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end05)

• Predictions—The availability of huge volumes and different, but highly relevant data, from within and outside healthcare, will make for better decisions especially about the future and grounded in predictive modeling.

We know the usual suspects in terms of where to concentrate and how analytics can contribute. Where to concentrate includes connecting the digital pipes, reducing the voltage drop in the translation of clinical guidelines into practice, measuring and improving the quality and safety of care, supporting healthcare reform, managing population health, and cracking the code on personalized medicine. How to do it with a panoply of analytics methods, theory, frameworks, and best practices has been described in this book. But there is something missing, and it has to do with spread and the bull’s eye.

Spread: Spread refers to the diffusion of a practice within and across organizations. For example, when a treatment approach has been proven to have a significant impact on patient outcomes through evidence-based research, one would assume that it would spread like wildfire through healthcare delivery systems. Or when an analytic approach has been teamed with a strategic goal and an implementation plan has been completed down to the last detail, one would expect a quick adoption and flawless implementation. Wrong on both counts! One of the most disappointing features of healthcare for those in the business of improving practice is the s-l-o-w adoption of innovations and low spread. The fact is that change is more complicated than good technology, strategy, incentives, and training manuals. This is discussed in a later section on making change happen.

The Bull’s-Eye: Most healthcare change strategies and analytics center on the 2-Ps: providers and payers. But there is a missing P—people. Most of the focus is on healthcare, but healthcare is a small contributor to achieving good health. And, the big, untapped resource for improving health is people. Only they can decide to see a doctor, get insurance, take their medication, follow orders, prevent illnesses through healthy behaviors, decide for themselves what treatments meet their preferences, sort out the providers that are best for them, and use their collective wisdom and power to change the system. The evidence is clear on this point. People control 40% of their health outcomes through their own behaviors.

Another part of the third P is regarding people as customers; that is, knowing them, understanding their needs, anticipating their needs, and serving them in the best definition of customer service.

The last part is understanding that people need to be supported to achieve good health. And it is much more than what happens in the doctor’s office.

The United States spends more per capita on healthcare, but it spends less than its peer wealthy countries on health.6

10/12/22, 1:52 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken… 38/45

(http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end06) Social services matter and, again, are more influential in determining health than is healthcare.

How can analytics serve these ends? Some examples:

• Attack the most dangerous epidemic facing the country—diabetes—by getting outside the box to use extra-industry data to find people with premorbid obesity.

• Engage people in measurement by addressing what matters. What matters is well-being, not whether the hospital is providing beta blockers to heart attack patients in a consistent way. Well-being should be measured, and every municipality and politician should be held accountable for the health of their communities just like they are for education, roads, and jobs.

• Give people dashboards that matter. These are not the self-service kind that instruct people how to fill out their insurance forms. People need compelling tools and active support. They need to be coached on their health journey. Other industries do this well, especially political campaigns.

• They need to be “known” such that services can be radically personalized to their needs. Analytics can provide the tools commonly used in customer analytics to know the person through microsegmentation, tailored messages that are delivered through preferred channels, and interventions that work. This is done in marketing to sell people things. It should be used in healthcare to connect with people to drive behavior change.

• Social supports are critical to health success. Knowing whether a person has financial, nutritional, transportation, and mobility needs can be quantified and are just as important as lab findings.

The bottom line on the way forward is to stay the course on the focus on providers and payers but to widen the aperture to include engagement and interventions with people. It is also important to ratchet up the spread of analytics that produce positive outcomes.

The Data Gold Rush The data gold rush in healthcare is on. It’s the wild, wild West to find and harvest it. We are told that the data produced in U.S. healthcare will soon be

counted in yottabytes,7 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end07) or a million trillion megabytes, or 1,000,000,000,000,000,000,000,000 bytes. We are told that creative and thoughtful extraction of all the healthcare big data is worth at least $300 billion a year. Where are all these data coming from, what value is being extracted from them, and where are the untapped opportunities?

Where Is It All Coming From?

The data come from a variety of sources:

• Transactions—The traditional sources of usable (structured) healthcare data come mostly from billing (claims) data.

• Electronic medical records—EMRs produce useful clinical data in mostly unstructured and semistructured data.

• Machine-emitted—Most of the yottabytes come from this source, which includes readings from medical sensors and “scrapings” from Web and social media sources including clickstream and social interaction data.

• Biometric devices—These are all the findings from medical measurements such as blood pressure readings and x-rays and other monitors of everything from steps taken (e.g., fitbit) to places visited (GPS).

• Research—Data on individuals from clinical trials, registries, and other sources.

• DNA sequencing—Genomic data to support personalized medicine are not widely available now but are on the verge of becoming accessible and reasonably priced.

What Value Is Being Extracted From Data?

It is very difficult to know what value is being extracted from these data sources for the purpose of healthcare analytics or for that matter how much money is being spent on healthcare analytics. Most of these data sources were developed for purposes other than healthcare analytics (as defined as providing insights for the enterprise to improve business and clinical outcomes). For example, claims data were developed for billing, and the use of them for understanding clinical analytics has been a stretch. EMRs are for the purpose of improving care and increasing communications among providers. They are not developed and implemented with the notion of combining the data with other sources to get a 360 degree view of the patient or to do comparative effectiveness studies. So, in a way, these data investments have already been made for specific other purposes, as have social media communications and purchases on the Internet. The use of these data for analytics to improve the business or to improve outcomes or develop new products is a secondary use. And the real value add of analytics may be to recombine and repurpose the data. (Later on we discuss the use of data collected primarily for healthcare analytics.)

Let’s focus on “big” data and “small” data and hypothesis-driven and hypothesis-free approaches. To start that discussion, let’s take a look at mining in another industry: gold.

There are 2,500 metric tons of gold produced annually.8 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end08) At the current price of

$1,300 per ounce this amounts to a $100 billion industry. It takes, on average, 30 tons of rock to produce one ounce of gold.9

(http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end09) Hence, the final product amounts to .000001042 of the rock that needs to be worked through to harvest it. There are also by-products of this extensive mining including the use of cyanide to extract it and huge open pits and large mounds of waste rock across the countryside where it is produced. The mining processes include huge investments in monster shovels and trucks to extract and transport rock to the plant and warehouse for processing and storing. Gold mining is hypothesis-driven; that is, mine rock in a specific

10/12/22, 1:52 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken… 39/45

place and in a specific way and you get gold. This is quite different from a hypothesis-free approach, which is to take all the rock and do a lot of tests on it to see whether there is anything in it of value.

Yotta-driven analytics in healthcare is mostly hypothesis-free, akin to analyzing the whole mountain and looking to discover “similarities” that may provide new understandings about the delivery of healthcare. The monster computing technology available today can enable seemingly limitless simulations to do this.

If we assume that analytics in all of its permutations in healthcare amounts to just 5% of healthcare spending, this computes to about $135 billion, which is pretty close to the gold mining industry. How much rock will it take to find the gold in healthcare? Will the conversion rate be +/- .000001042?

Some of the gold from the healthcare data rush is palpable and it is “small.” The integration of genomic data with clinical data could lead to answers to important questions, such as whether a certain chromosomal variation is related to a disease, which could then fuel individually tailored treatments. For example, Tamoxifen has been an effective drug for the treatment of breast cancer. On average, about 80% of patients benefit from it. The potential with personalized treatment is to become 100% effective in 80% of patients because genetic markers can improve the knowledge of who does and does not benefit from the treatment. There are many instances of “small” hypothesis-driven data that can have a precise impact on business and health outcomes. Other rock in the yotta may not be as clearly useful. For example, much of the yotta is comprised of data emitted from machines, and much more research needs to be conducted to home in on likely ways it can contribute.

Untapped Opportunities

There are two types of data missing from the previous list. These data do not necessarily add a lot to the yotta stats. They are “small” and have specific and targeted purposes. These include extra industry personal data and people-generated data.

Extra Industry Personal Data

The world is full of relevant data and a lot of it resides outside of healthcare. External data can address specific healthcare issues, for example, to change people’s behavior, ranging from marketing to early detection of diseases. These data come from privately aggregated and publicly available databases on a wide range of personal attributes that can define microsegments that can be precisely targeted with specific interventions to improve health. For example, data on height and weight are available from external sources (and not easily collected or extracted from usual healthcare data) and can be used to calculate the body mass index (BMI) to determine premorbid obesity. Additionally, when personal data are integrated with medical data and in combination with the right channel—especially mobile—it can produce a much better identification of high-risk patients, with more effective interventions mapped to their specific needs, and include closer monitoring over time.

People Generated Data

Another source of untapped data is people. This is another type of “small data” with big potential benefits. Most of the data sources listed previously do not involve the active participation of people. The real potential lies in gathering much more relevant data from individuals with their consent and engendering their partnership to engage in data-sharing activities that help them improve their life. After all, people know more about their own health and illnesses and can monitor it better than any doctor could possibly hope to do. There is much more to be learned from a person’s head than from their data streams. There are indications that this is happening without, and perhaps in spite of, the active strategies of traditional healthcare. For example, networks of patients with the same condition are sharing data and creating large databases that are beginning to approximate crowd- sourced clinical outcomes research. For example, as of the end of 2011, PatientsLikeMe had more than 120,000 patients in 500 different condition groups; ACOR (Association of Cancer Online Resources) had more than 100,000 patients in 127 cancer support groups; 23andMe has more than 100,000 members in their genomic database. People also engage in their own data sharing through mobile and social media. And people have been responsive to surveys when the purpose is big (like polling in a presidential campaign) and when the rewards for participation are adequate.

Conclusion

A mountain of data is available for analytics in healthcare. Much of it was collected for another function and may be repurposed. Some of it is really big and has unknown uses but is intriguing, and the technology may be able to find the gold although the conversion rate may be infinitesimally small. Some of it is small and can have immediate applications to produce value. And some that is potentially very valuable and comes directly form people is not included in the count and is not collected. Certainly, healthcare lags other industries in its use of big data because of the challenges with complex and unstructured data, the reluctance to use external data, data integration issues, and concerns about patient confidentiality. And IT folks say there is enough unused healthcare industry data to keep them busy for a very long time. Threading the needle for the most productive use of data, whether big or small, hypothesis driven or free, depends on analytics making the case that it is worth the investment and an innovation worth adopting.

Technology, The Great Enabler Technology is a great enabler for analytics to contribute to the success of an enterprise in the sense that it supplies the computing means and capabilities (hardware, software, and know-how) to solve problems related to data. Moore’s law predicted that chip performance would double every two years, which would increase processing speed, memory capacity, sensors, and even the pixels in digital cameras proportionately. For example, comparing the IBM PC released in August 1981 with the Apple iPhone 4 released in June 2010, the CPU clock speed of the PC was 4.77MHz compared to iPhone at 1GHz; the processor instruction size was 16 bits for the PC and 128 bits for the iPhone; the storage capacity of the PC was 160KB and that

of the iPhone (base model) was 16GB; and the installed memory (RAM) was 64KB for the PC and 512MB for the iPhone.10

(http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end10) Additionally, the list price on release of the PC was $3,000 (or about $7,500 adjusted for inflation) and the iPhone was $199, or about 2.5% of the cost of the original PC. This exponential growth in computing performance has driven the impact of digital devices from computers to household appliances in every segment of the world economy.

10/12/22, 1:52 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken… 40/45

Has this exponential increase in computing performance had a concomitant increase in information and analytics to improve healthcare? Far from it. One could say that the technology has been ahead of the capacity of organizations to absorb it. It’s not the technology that accounts for a slow take-up of analytics, it’s the sociology. The hope for technology adoption reminds me of Terrence Mann’s prediction about building a ball field in the cornfields of Iowa in the movie Field of Dreams: “If you build it, he will come. Ray, people will come, Ray. They’ll come to Iowa for reasons they can’t even fathom. They’ll turn up your driveway not knowing for sure why they’re doing it.”

In healthcare, organizations do not show up in the driveway to buy technology and do not make large investment decisions without knowing exactly what the ROI is. Technology adoption for analytics in healthcare has not been as rapid as that in science and other industries for a variety of reasons. Healthcare has a cultural underpinning of “do no harm,” and it shows in its hesitant approach to change. Why healthcare technology, and analytics innovations in particular, suffer a slow pace of adoption is addressed in the next section on making change happen.

The technology breakthroughs available to healthcare are awesome and can make analytics quicker, cheaper, and smarter. A few are discussed briefly in the following sections.

NoSQL (Not Only SQL)

These databases are an alternative to traditional, relational databases and are especially suited for unstructured big data, Web 2.0, and mobile applications. It uses open source software that supports distributed processing. It scales “out” to the cloud, rather than “up” with more servers. It has fewer data model restrictions than relational databases management systems, which allows more agile changes and less need for database administrators. It can use low cost commodity hardware. The bottom line is that it is faster and much cheaper. Examples of popular NoSQL databases include Cassandra, Hadoop, and BigTable. Companies that use it include Facebook, Netflix, LinkedIn, and Twitter. For more information see the NoSQL

website, which touts itself as “your ultimate guide to the non-relational universe.”11

(http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end11)

High Performance Computing (HPC) Through the Cloud

High performance computing (HPC) allows users to solve complex science, engineering, and business problems using applications that require high bandwidth, low latency networking, and very high compute capabilities. This is the computing capability needed for mining mountains of data. This capacity can be provided by dedicated computer clusters or by cloud clusters. Dedicated, custom-built, supercomputer infrastructure requires significant capital investments, long procurement times, long queues, and extensive database management. Buying HPC services from the cloud provides definite cost advantages, short lead teams, access to the scale required for a given project, and on-demand capacity. An example of such an

offering is from Amazon Web Services called Cluster Compute Instances.12 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end12) In healthcare, the biopharma sector uses HPC for genome analysis. Other industries, including oil and gas, financial services, and manufacturing, use it for modeling.

Machine Learning

The idea that machines could replace humans for certain functions has been around a long time. And it certainly has become commonplace in industries such as automotive with robots on the assembly line. But can the machines actually “learn” and improve functioning on their own beyond being explicitly programmed? There are good examples of this with Google Search and Amazon purchasing recommendations, and with voice and facial recognition applications. In healthcare, IBM demonstrated a compelling use of machine learning (and natural language processing and predictive analytics) with its Watson technology by beating two grand champions on the Jeopardy! TV quiz show. IBM is now moving beyond quiz shows and working on healthcare solutions, mostly in the area of differential diagnosis. One of the institutions it has partnered with is Memorial Sloan-Kettering

Cancer Center.13 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end13) The goal according to Sloan-Kettering is to have the technology gather and assimilate information from the research literature and from the Center’s clinical experience documented in its medical records and other

files to “bring up-to-date knowledge to the bedside of every cancer patient.”14 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end14)

Watson might be able to do this through its capabilities to read and understand language, interact with humans, remember everything, and provide answers to real-time questions. How the information will be delivered to the physician, how it might transform the practice of medicine, and whether physicians will embrace the technology are all important, open questions. It is certainly a very bright star illuminating and guiding the emerging field of machine-enabled clinical decision support.

Clicks and Mobile

The Internet has transformed the way businesses communicate, market, do commerce with customers, and collect data about them. In retail, clicks are challenging the bricks. What could be more indicative of shifting paradigms than the collapse of the structures in which people do business (stores). One example is the capability to do randomized trials, or A/B testing, of alternative Web site features—for example, how to get the most contributions during a political campaign—on large samples and virtually instantaneously. Another example is Web page “scraping” in which all types of data about people’s Web wanderings are turned into ratings about their suitability for a job, a loan, and a date.

More than half of the adult population in the United States have smartphones.15 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end15)

Facebook has more than 1 billion monthly users.16 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end16) The hot combination of these two popular technologies produces a platform for easy, convenient, and quick communications that also enable e-commerce, uber-targeted marketing, location monitoring, and much more. An opportunity going forward in healthcare is to create closer relationships with people to help them get healthier by tapping into data that are freely exchanged and by supporting the continual, fast evolution of new applications to support health.

10/12/22, 1:52 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken… 41/45

The Two Faces of Enabling

Technology can enable analytics in two ways. The positive face of enabling is seen in its tremendous success story in increasing computing capacity with hardware (speed, memory, storage, access, etc.) and with software (to manage all the data and make sense of it). But technology is mostly content-agnostic. It really does not care about the content or use of the data. It is a template for manipulating it. And often, it perpetuates GIGO, or garbage in/garbage out. That is, if the data, models, and assumptions that are foisted on the computer by humans are wrong, the computer cannot correct them. And now, with the capacity to process more data more quickly, there is the possibility of more garbage.

The flip side of enabling is to keep people from experiencing and learning about challenges and their consequences by unwittingly helping them, but in the wrong way. This may serve to protect the individual from harm but actually exacerbates the discovery of new solutions. Technology can be so seductive to analysts that it blinds them from keeping their eyes on the prize. The prize is changing behavior to improve health and reduce costs. And with most of these challenges in healthcare, there is an adequate supply of technology. What’s missing is the time and attention it takes to make change happen with people, not machines.

Making Change Happen With the scorching, burning platform for healthcare transformation; the availability of huge volumes and diverse types of data; technologies to make computing quicker, better, and cheaper; and hype about the art of the possible with big data; one might ask why the uptake of healthcare analytics is so low in comparison to its potential and relative to the performance of other industries? In addition to the usual attributions including the need to perfectly digitize the industry, there are industry context barriers that deter innovation, including lack of good coordination, physician autonomy, convoluted market dynamics, multiple vested and powerful interests, and a pervasive, risk averse culture. But, I would like to concentrate on two possible answers and solutions that analysts can and should address: The technology of change is underrated, and analysts must change their understanding of their role and contribution.

Technology of Change

Making change happen is not as rational as we analysts would hope. We assume that if we demonstrate the ROI for an innovation and put it in a good business plan that decision makers will burn a path to our door to implement our recommendations. It’s hard enough to do this for conventional business functions but harder for analytics innovations. Making change happen is much more complicated and subtle, and much less about numbers than about people.

Healthcare works really hard to change people’s behavior. It borrows from aviation to have pre- (flight) surgical checklists to reduce mistakes like wrong leg amputations, it issues alerts (continuously), for example, when the wrong medication is ordered on the CPOE (Computerized Physician Order Entry) system, it measures and publicly reports on the performance it wants because it believes what is measured is managed well, and it conducts a lot of research to get the evidence on what works and what doesn’t to improve the knowledge base and shape the delivery of care.

Most strategies to get people to change the way they do things often boil down to three approaches: Please do it and I will teach you how; you must do it and I will punish you if you don’t; and do it well and I will pay you a (modest) reward for doing so. These are all rational approaches and make sense and are a part of the armamentarium of change....And they are insufficient.

What we are trying to do with analytics is embed them in the daily operations of an organization such that it becomes the normative way of doing things. More than that we want the analytics to be successful and produce the intended results on an assured, repeatable basis. Eventually we want the analytics innovation to give way to its own reinvention as it adapts to the needs of an organization. This is a long journey and cannot be short circuited. It involves mastering six stages:

1. Ideas—Ideas are the starting point and backbone of innovations. President John F. Kennedy exalted ideas. He said, “A man may die, nations may rise and fall, but an idea lives on.”

2. Design—Ideas need to be converted into a theory of action on how it will accomplish a goal.

3. Decision—This is when the design is proposed to decision makers, and it receives an up, down, or delay action. Most ideas do not pass the test for adoption. Machiavelli said, “There is nothing more difficult to plan, more doubtful of success, nor more dangerous to manage than the creation of a new order of things.” More on this later.

4. Implementation—Implementation is about following the design rule book in carrying out all the required process steps. Goethe noted in the eighteenth century, “To put your ideas into action is the most difficult thing in the world.”

5. Evaluation—Stakeholders need to understand performance to make adjustments in the design and operations so that it becomes institutionalized, altered to fit the changing needs of the organization, or discontinued. More often than not, ideas get stuck in the pathway and do not live to fight the next stage. And implementations of complex programs suffer from innumerable snags in delivery and most fail.

6. Reinvention—Reinvention is important and is not a failure of intent to implement a plan “as written.” For example, one might think that clinicians would adopt clinical guidelines developed by their professional organizations as is. But success might depend on not getting the full loaf and instead adopting a local version that works in a particular context. Hopefully, all innovations evolve and change through a learning and improvement process.

Analytics cannot be responsible for all elements of the innovation pathway. After all that is what operations folks and c-suite leadership are about. But what analysts must do is master the first three stages including ideas, design, and decision. Analysts must know how to get their projects known and funded.

The technology of adoption is complex and multifaceted. Adoption is about making a decision to activate an innovation into practice. The process of adoption is largely about collecting, processing, and evaluating information to understand the innovation and to reduce uncertainty in relation to its pros and cons. A useful model based on the works of multiple scholars of making change happen is displayed in Figure C.1

10/12/22, 1:52 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken… 42/45

(http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01fig01) . The model includes six domains consisting of 18 factors which need to be addressed for the successful adoption of an innovation:

(Source: Dwight McNeill, A Framework for Applying Analytics in Healthcare: What Can Be Learned from the Best Practices in Retail, Banking,

Politics, and Sports.17 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end17) )

Figure C.1 Innovation adoption factors model

• Innovation receptivity—This includes the important consideration of receptive context, that is, whether the land (organization) is fertile, arable, and moist enough to grow seed. To what extent is the organization receptive to new ideas? Does it nurture new ideas or penalize the troublemakers who question the status quo? Are there survival pressures that require change such as the drivers stated earlier? What are the existing norms and beliefs about change in general? Is the culture rooted in values of “better to be safe than sorry” and “do no harm” and conservative in taking on risks?

• Idea maturation—This includes whether the idea is articulated sufficiently and is ready for prime time. Has it confronted confirmation bias and looked beyond “seeing only the things we are looking for”? Is it clear about the cause and effect theory? If this is not clear, and what the steps are, the argument can shift to ideological and political issues.

• Urgency and timing—Is there a burning platform for change that the idea is associated with and is there a window of opportunity within which all the stars must get aligned?

• Innovation attributes—The innovation attributes domain is important and includes the key factors of the relative advantage of the innovation, its compatibility with existing processes and attitudes, its complexity, and its “trialability.” Of most importance is whether the innovation is judged to be better than an alternative. What is its ROI? Is it compatible and fits in with the way the organization does things? How complex is it to understand and use? Can it be given a test drive before a full scale roll-out?

• Organization capabilities—Who is making the decision to adopt—an individual, team, or ultimately the users who can support or torpedo it? Is the program component of the analytics innovation capable of making it work? For example, analytics can precisely segment people into their risk for morbid obesity. But is there a compelling theory about how program staff actually change the behaviors of those identified? Similarly, does the organization have the people skills and technologies to implement the innovation successfully?

• Readiness—This includes evidence from small scale testing of an innovation to show its impact and communications to position the innovation favorably such that those who are expected to deploy it are ready and willing. And last but not least is leadership ready and willing to support it? On the last point: Leadership at the executive level is about creating the vision and executing a strategy to “get people from where they are to where they have not been,” in the words of Henry Kissinger. Succeeding at innovation requires leadership that creates an environment that welcomes and encourages new ideas and change, cuts through the confusion and uncertainty to be decisive, and then leads the organization to execute flawlessly. Ideally, the organization is a learning system that is constantly updating its mental models and making improvements. Without this executive leadership, the likelihood of adoption of new innovations is diminished.

In summary, there are six stages of the innovation pathway to success. For one of these stages, decision, there are six domains and 18 factors. For the other stages, there is the same degree of complexity along with clear frameworks for addressing it. So, there is a technology of change that is well articulated.

Atul Gawande, in his essay in the New Yorker, “Slow Ideas,”18 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end18) boils this all down

to a simple suggestion. He reiterates the wisdom of Everett Rogers19 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end19) from 50 years ago that “Diffusion is essentially a social process through which people talking to people spread an innovation.” Gawande presents a case study on how to get hospitals and birth attendants in India to carry out a few of the tasks required for safer childbirth, such as washing hands, keeping the baby warm, and taking vital signs properly. He asked a nurse why she changed her behavior and how it became something she did “day in and day out, even when no one is watching.” She said that she did not listen to the teacher at first because of her heavy workload, but that after multiple visits with the teacher/coach she began to listen and change. Asked why, she said the teacher was “nice.” “It wasn’t like talking to someone who was trying to find mistakes. ...It was like talking to a friend.” He poses the question, “So, what about just working with healthcare workers, one by one to do just that?” Sounds corny? A fundamental, yet overlooked, cornerstone of making change happen is people-to-people communications, done often, and with great respect toward the individual’s existing norms and beliefs.

10/12/22, 1:52 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken… 43/45

“Be the Change...” If change is not optimal in healthcare and analytics is not precipitating it, then something has got to give. In this section I address the field of analytics professional identity, salvation by the Chief Analytics Officer (CAO), and our individual role to “be the change.”

Identity Crisis

Who are we (as analysts) and what do we do? What is our identity? Identity is defined as the state or fact of remaining the same one under varying aspects or conditions. How do analysts define themselves? The words that come to mind most often include quant, stathead, and data scientist. How we define ourselves reflects our purpose and what work we do.

What do we do and how do we contribute (functions)?

• Create infrastructure for generating and analyzing data.

• Produce platforms for data sharing.

• Create one version of the truth.

• Adopt the best information technologies.

• Keep information infrastructure costs low.

• Produce standard reports and ad hoc queries.

• Produce products that can support decision making.

• Execute information management strategies.

• Work collaboratively with leaders across the enterprise to support their achievement of strategic priorities.

• Provide insights that are valuable to the business.

• Contribute to the growth of the business by demonstrating a contribution to the earnings per share.

All of these are relevant to what analysts do. Those at the top of the list are more common today. Those at the bottom are emerging as the field advances in its development and focuses on the ends (outcomes) more than the means.

What do we need to know to contribute to purpose (competencies)?

• Data

• Computing and analytics methods and technologies

• The nature of the business and knowledge about its customers

• How analytics creates value

• How to negotiate change in an organization

• How to be a leader

The first two competencies are what we most often think of when we think of analytics. Our contributions get more significant as we go down the list and ratchet up our engagement in making the business work better and grow.

So, our identity is changing, hopefully, as we advance in the maturity of our profession and change the scope of our work from doing things with data to using data to change the organization for the better.

But this is an overwhelming list of functions and competencies. No one person can do it all. It requires a village (team) to have all the skills and accomplish all the functions. And some suggest that a new chief is needed.

Salvation from the Chief Analytics Officer (CAO)?

CAO stands for chief analytics officer. (It also stands for chief administrative officer.) The role is new for the C-suite. Michael Bloomberg, the data-

driven mayor of New York City, in his last State of the City Address,20 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end20) appointed the city’s first ever CAO, Michael Flowers, to improve the way all agencies share information and to make the data available to the public so that the community can hold the city accountable. Flowers used to be the city’s analytics director. It’s not clear why there was a title change. The job description sounds better in the previous job. “Mr. Flowers leads a team of data scientists in analyzing city data from over 20 city agencies to allocate

its resources quickly and efficiently to prevent fire, crime, safety hazards, and unhealthy conditions.”21

(http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end21)

Chiefs are becoming very popular. Forbes lists new C-suite titles22 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end22) including chief Internet evangelist, chief happiness officer, chief privacy officer, chief digital officer, chief knowledge officer, and chief customer officer among others. In healthcare, the new CAO role joins forces with other information leaders including the CIO (information), CDO (data), CIO (innovation), CMIO (medical informatics), and CNIO (nursing informatics). I bet there are more to come.

I guess the reason for chiefs is to bring visibility to the function, get the ear of the CEO, collaborate with peer chiefs for the good of the enterprise, provide better management oversight, and be accountable for results. All good things, and it’s important that analytics is recognized as an important function along with the dozens of others. It is good to have the executive talent, but leadership is not just for the few chiefs. It’s for all of us.

10/12/22, 1:52 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken… 44/45

Remake Ourselves

Mahatma Ghandi said “Be the change that you wish to see in the world.” We need to expand our technical and people skills to increase the utility of analytics in healthcare. We need to work locally and make teams work better through communications and collaboration and dedication to a common goal. We need to focus on the immediate tasks at hand, such as working through an algorithm or building a database and also be sure there is a receptor site to absorb our work. We need to visualize how analytics improves business and society. Ultimately, we need to lead by our own example.

Notes 1 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end01a) . James Manyika et al. Big Data: The Next Frontier for

Innovation, Competition, and Productivity. McKinsey Global Institute, May 2011. http://www.mckinsey.com/insights/mgi/research/technology_and_innovation/big_data_the_next_frontier_for_innovation (http://www.mckinsey.com/insights/mgi/research/technology_and_innovation/big_data_the_next_frontier_for_innovation)

2 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end02a) . Atul Gawande, “Slow Ideas,” New Yorker (July 29, 2013), p. 45.

3 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end03a) . Dwight McNeill, A Framework for Applying Analytics in Healthcare: What Can Be Learned from the Best Practices in Retail, Banking, Politics, and Sports (Upper Saddle River, NJ: FT Press, 2013).

4 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end04a) . Ibid. p.31.

5 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end05a) . Suzy Hansen, “What Can Mississippi Learn from Iran,” New York Times, July 27, 2012.

6 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end06a) . Elizabeth Bradley and Lauren Taylor, “To Fix Health, Help the Poor,” New York Times, www.nytimes.com/2011/12/09/opinion/to-fix-health-care-help-the-poor.html??_r=0 (http://www.nytimes.com/2011/12/09/opinion/to-fix-health-care-help-the-poor.html??_r=0)

7 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end07a) . P. Cerrato, “Is Population Health Management the Latest Health IT Fad?” Information Week (July 31, 2012) retrieved December 12, 2012, http://www.informationweek (http://www.informationweek) .

8 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end08a) . “All the World’s Gold,” http://www.numbersleuth.org/worlds-gold/ (http://www.numbersleuth.org/worlds-gold/) .

9 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end09a) . “Behind Gold’s Glitter: Torn Lands and Pointed Questions,” New York Times, June 14, 2010, http://www.nytimes.com/2005/10/24/international/24GOLD.html?pagewanted=all (http://www.nytimes.com/2005/10/24/international/24GOLD.html?pagewanted=all) .

10 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end10a) . Alfred Poor, “30 Years Later: An Unfair Comparison Between an IBM PC and an Apple iPhone 4,” HP.com, December 4, 2012, http://h30565.www3.hp.com/t5/Feature-Articles/30-Years-Later- an-Unfair-Comparison-between-an-IBM-PC-and-an/ba-p/2662 (http://h30565.www3.hp.com/t5/Feature-Articles/30-Years-Later-an-Unfair-

Comparison-between-an-IBM-PC-and-an/ba-p/2662) .

11 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end11a) . NoSQL.org, http://nosql-database.org/ (http://nosql-

database.org/) .

12 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end12a) . Amazon Web Services, What’s New?, “Announcing High Memory Cluster Instances for Amazon EC2,” http://aws.amazon.com/about-aws/whats-new/2013/01/21/announcing-high-mem-cluster- instances-for-amazon-ec2/ (http://aws.amazon.com/about-aws/whats-new/2013/01/21/announcing-high-mem-cluster-instances-for-amazon-ec2/) .

13 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end13a) . IBM, “Memorial Sloan-Kettering Cancer Center, IBM to Collaborate in Applying Watson Technology to Help Oncologists,” March 22, 2012, www-03.ibm.com/press/us/en/pressrelease/37235.wss (http://www-03.ibm.com/press/us/en/pressrelease/37235.wss) .

14 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end14a) . Larry Norton, “How Collaboration Between IBM and Memorial Sloan-Kettering Taps the Wisdom of Physicians,” Huffington Post, Blog, March 29, 2012. www.huffingtonpost.com/dr-larry-norton (http://www.huffingtonpost.com/dr-larry-norton) .

15 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end15a) . Mark Rogowsky, “More Than Half of Us Have Smartphones Giving Apple and Google Much to Smile About,” Forbes.com, June 6, 2013, http://www.forbes.com/sites/markrogowsky/2013/06/06/more-than-half-of-us-have-smartphones-giving-apple-and-google-much- to-smile-about/ (http://www.forbes.com/sites/markrogowsky/2013/06/06/more-than-half-of-us-have-smartphones-giving-apple-and-google-much-to-

smile-about/) .

16 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end16a) . Donna Tam, “Facebook by the Numbers:1.06 Billion Monthly Active Users,” CNet.com, http://news.cnet.com/8301-1023_3-57566550-93/facebook-by-the-numbers-1.06-billion-monthly- active-users/ (http://news.cnet.com/8301-1023_3-57566550-93/facebook-by-the-numbers-1.06-billion-monthly-active-users/) .

17 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end17a) . Dwight McNeill, A Framework for Applying Analytics in Healthcare: What Can Be Learned from the Best Practices in Retail, Banking, Politics, and Sports (Upper Saddle River, NJ: FT Press, 2013).

18 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end18a) . Atul Gawande, “Slow Ideas,” New Yorker, July 29, 2013.

19 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end19a) . Everett Rogers, Diffusion of Innovations, first edition (New York: Free Press, 1961).

20 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end20a) . “Mayor Bloomberg Delivers 2013 State of the City Address,” NYC.gov, Press Release, February 14, 2013, online at http://www.nyc.gov/portal/site/nycgov/menuitem.c0935b9a57bb4ef3daf2f1c701c789a0/index.jsp?

10/12/22, 1:52 PM Print

https://content.uagc.edu/print/McNeill.2947.17.1?sections=part04,ch17,ch18,ch19,ch20,ch21,ch22,ch23,ch24,ch25,app01&content=all&clientToken… 45/45

pageID=mayor_press_release&catID=1194&doc_name=http%3A%2F%2Fwww.nyc.gov%2Fhtml%2Fom%2Fhtml%2F2013a%2Fpr063- 13.html&cc=unused1978&rc=1194&ndi=1 (http://www.nyc.gov/portal/site/nycgov/menuitem.c0935b9a57bb4ef3daf2f1c701c789a0/index.jsp?

pageID=mayor_press_release&catID=1194&doc_name=http%3A%2F%2Fwww.nyc.gov%2Fhtml%2Fom%2Fhtml%2F2013a%2Fpr063-

13.html&cc=unused1978&rc=1194&ndi=1) .

21 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end21a) . “Michael Flowers,” Champions of Change, The White House, http://www.whitehouse.gov/champions/local-innovation/michael-flowers (http://www.whitehouse.gov/champions/local-

innovation/michael-flowers) .

22 (http://content.thuzelearning.com/books/McNeill.2947.17.1/sections/app01#app01end22a) . “The New C-Suite Titles,” Forbes.com, http://www.forbes.com/pictures/fhgl45eglf/chief-internet-evangelist/ (http://www.forbes.com/pictures/fhgl45eglf/chief-internet-evangelist/) .